mm if i's|s:vi>3& Y*/ MJiw ''^J feXa %P iMti %^l ^W X Mmm Qi '."^THE JOURNAL @NJ THE « @F RsESlA.^CHJ m. Established in 1 962 Edited by WILLIAM HOVANITZ Volume 1 1962 - 1963 f Published at 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA is sponsored by THE LEPIDOPTERA FOUNDATION CONTENTS Volume 1 Number 1 August, 1962 Editorial 1 Variation in the silvering of Argynnis (Speyena) cdlippe in the interior mountain area of south central California . . . Oscar Elton Sette 3 The effect of various food plants on survival and growth rate of Pieris .... William Hovanitz and Vincent C. S. Chang 21 General characteristics of the movements of Vanessa cardui (L.) J. W. Tffden 43 Three factors affecting larval choice of food plant William Hovanitz and Vincent C. S. Chang 51 The generic, specific and lower category names of the Nearctic butterflies. Part 1 •— The genus Pieris . . Paddy McHenry 63 The distribution of the species of the genus Pieris in North America William Hovanitz 73 Did the caterpillar exterminate the giant reptile? .... S. E. Flanders 85 Further evidence of the distribution of some boreal Lepidoptera in the Sierra Nevada Clyde Eriksen 89 Argynnis and Speyer ia . William Hovanitz 94 Volume 1 Number 2 January, 1963 Composition and relative abundance in a temperate zone butterfly fauna . . . Thomas C. Emmel and John F. Emmel 97 The Argynnis populations of the Sand Creek area, Klamath Co., Oregon, Part I .... J. W. Tilden 109 Caterpillar Versus Dinosaur? . . . Theodore H. Eaton, Jr. 114 Geographical distribution and variation of the Genus Argynnis I. Introduction II. Argynnis idalia William Hovanitz 117 The relation of Pieris virginiensis Edw. to Pieris napi L. Species formation in Pieris? ..... Wiliam Hovanitz 124 The male genitalia of some Co lias species . Bjorn Petersen 135 The effect of hybridization of host-plant strains on growth rate and mortality of Pieris rapae . WiUiam Hovanitz and Vincent C. S. Chang 157 Notice-editorial 162 Change of food plant preference by larvae of Pieris rapae controlled by strain selection, and the inheritance of this trait William Hovanitz and Vincent C. S. Chang 163 Volume 1 Number 3 March, 1963 Selection of allyl isothiocyanate by larvae of Pieris rapae and the inheritance of this trait William Hovanitz and Vincent C. S. Chang 169 Biology of the Ceanothus stem-gall moth, Periploca ceanothiella (Cosens), with consideration of its control. J. Alex Munro 183 Larval food-plant records for six western Papilios, .... John F. Emmel and ThtMnas C. Emmel 191 CoUas philodice in Chiapa^ Mexico . . Thomas C. Emmel 194 Notes on the early stages of two California geometrids John Adams Comstock 195 Geographical distribution and variation of the genus Argynnis III. Argynnis diana William Hovanitz 201 The generic, specific and lower category names of the nearctic butterflies. Part 2 — The Genus Colias . Paddy McHenry 209 A standard method for mounting whole adult lepidoptera on slides utilizing polystyrene plastic . . Charles L. Hogue 223 Volume 1 Number A May, 1963 Techniques in the study of population structure in Philotes sonorensis Rudolf H. T. Mattoni and Marvin S. B. Seiger 237 Early stages of a southern California Geometrid moth, Drepanulafrix hulsti hulsti (Dyar) John Adams Comstock 245 The effeaiveness of different isothiocyanates on attraaing larvae of Pieris rapae William Hovanitz, Vincent C. S. Chang and Gerald Honch 249 The origin of a sympatric species in Colias through the aid of natural hybridization William Hovanitz 261 A method for breeding Pieris napi and Pieris bryoniae Bjorn Petersen 275 An analysis of the North American species of the genus Callophrys J. w. Tilden 281 THE LEPIDOPTERA FOUNDATION 301 Notice 302 Volume 1 Number 1 August, 1962 Established in 1 962 Published at 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Edited by WILLIAM HOVANITZ WITH EMPHASIS ON ENVIRONMENTALLY AND GENE- TICALLY INDUCED VARIATION, population analysis, evolution, phylogenetic taxonomy, zoogeography, comparative morphology, ecol- ogy, geographical variation, speciation, physiology, etc. In short, quality tuork on any aspect of research on the Lepidoptera. THE PURPOSE OE THE JOURNAL is to combine in one source the work in this field for the aid of students of this group of insects in a way not at present available. The JOURNAL will attempt to publish primarily only critical and complete papers of an analytical nature, though there will be a limited section devoted to shorter papers and notes. RATES: $ 8.00 per volume, personal subscription. $12.00 per volume, institutional subscription. Benefactor subscriptions are encouraged. These may be any amount over those above. All amounts are in U.S. dollars, payable in the U.S.A. AUTHORS ARE REQUESTED to refer to the Journal as an example of the form to be used in preparing their manuscripts. Fifty separates will be supplied to authors free; reprints will be sold at printer’s rates if ordered at time galley proofs are returned. If proofs are not returned promptly, the editor reserves the right to withhold publication, or to proceed with no responsibility. All issues of the Journal will be copyrighted; in submitting a paper each author agrees that the material has not been published elsewhere. MANUSCRIPTS AND SUBSCRIPTIONS should be mailed to the address noted above. BENEFACTORS Individuals or institutions who by reason of their interest in the welfare or research on the Lepidoptera, and in its publication, who wish to contribute beyond the cost of a subscription, will be listed here. BIO-METAL ASSOCIATES THOMAS C. EMMEL ANONYMOUS DONOR ANONYMOUS DONOR 1(1):1,1962 Journal of Research on the Lepidoptera 1140 W. Orange Grove Ave., Arcadia, California, U,S.A. Copyright 1962 EDITORIAL Starting a new journal is not an easy matter, nor is it a matter free of controversy. This is especially the case where a new journal, as this one, is started without the support of an existing society. This journal is intended as a meeting ground for papers on all aspects of the biology of the Lepidoptera, moths and butterflies alike. Too many entomolog- ical journals don’t offer publication of articles of the length necessary to present sound data and Completeness of effort. The Lepidoptera are a group of insects second to none in the scope of our knowledge about them: work in this fieM has reached the higher levels of learning and a journal is required to keep pace with this knowledge at the same level. A journal with a wide scope is our aim — narrow in the taxonomic sense but wide in the biological sense. The journal will attempt to publish primarily only critical and complete papers of an analytical na- ture, though there will be a limited section devoted to shorter papers and notes. There will be emphasis on environmentally and genetically induced variation, population analysis, evolution, phylogenetic taxon- omy, zoo-geography, comparative morphology, ecology, geographical variation, speciation, physiology, etc. It has been very gratifying to receive so many subscriptions on the basis of merely a brochure; every effort will be made to keep the journal up to the standard expected of us. The response in so short a time has served to show those who have generously provided the finances that a journal of this type has been needed and will fill a desirable niche. We are happy to publish in this first issue a number of papers which the authors were able to get ready on "short notice". There is a preponderance on one taxonomic group, but with different points in mind. With this first issue out, we offer the pages of this journal to all with papers to publish on the Lepidoptera that fit the intent of the journal. A wider range of papers will come in future issues. 1 v:,.-;:, -■■■■■• fe l« •'■■'v-’V' ' ■' I,;':?-*' :slly ■■,;■■■:' -V. 'V-..' - -'.r ' ' ' ":. •i^-' . ■' ;^ :,;■ ^ .,.:;v. ... r; v^;4 ■ 4' - •-■■ *v4: - 1 ' -.: .;V ^ ^ ^ '■ ' v<j: ' ■:.;-q - • ,■ ’■. : . ’"V', 'U;:: ■ '• v;:'^ ' '4 _ ■'■'-■■ 4.;vvK'^j^ ■■ •:.■■■•; 4.' :-i v4-:;.,\:;.. ;-:' .,; ^ 44' ,;!4.^4| , ; . ,' ..' ,. ■ ;:v.M ^4ir S''’^'"' '* • 'V' ' •‘'-'i > * <■ i^‘ . ‘ '■‘-'v' * ' <’ “•^4 - ' ’ ' " '. l(l):3-20, 1962 Journal of Research on the Lepidoptera 1140 W. Orange Grove Ave,, Arcadia, California, U.S.A. Copyright 1962 VARIATION IN THE SILVERING OF ARGYNNIS (SPEYERIA) CALLIPPE IN THE INTERIOR MOUNTAIN AREA OF SOUTH CENTRAL CALIFORNIA OSCAR ELTON SETTE 23643 Arbor Avenue, Los Altos, California In a comprehensive study of the geographic variation of Argynnis (Speyeria) callippe Boisduval, based on examination of about 2750 specimens from many localities well distributed through out the range of the species in California, Hovanitz (1943) recog- nized several main divisions of this species complex. Among them are the *'South Coast Range Population” extending from the San Fran- cisco Bay area southward through the Coast Ranges into Lower Calif- ornia and the "Western Sierra Nevada Populations” extending in a band along the western Sierra Nevada, at moderate elevations from the region west of Lassen Peak southward to about the Greenhorn Mountains at the southern end of the Sierra Nevada. He found that these two population groups are connected through a "Southern Zone of Intergradation” by a "series of steps across the Piute Mountains, the Tehachapi Mountains and the Sierra Madre Range.”’ He characterizes the populations in the Southern Zone of Inter- gradation as follows: "From the Santa Monica Mountains ‘ on the coast, where the typical callippe of this (the South Coast Range) region lives, it is found that in going inland (Charlton Flat, Mint Canyon and ’Ridge Route’) the lightness of all colors increases. The butterfly becomes smaller and the light colored band and spots on the upper surface of the wings tend to become obliterated, leaving a more uniformly colored wing surface such as is present in the western Sierra Nevada gradient. However, the yellow- brown color is very much lighter than in the latter and the band on the under side of the hind wings is still yellow; the spots are still always fully silvered. The tendency toward these conditions is the more marked the farther from the coast and the farther into the Tehachapi Range the populations exist. In the Tehachapi Range, the butterflies are very lightly colored and the band on the upper surface of the wings is rare; the spots are still silvered. At Havilah, Piute Mountains, the population consists of some silvered, some unsilvered and some intermediate spotted IThis range is not named on many maps. It lies parallel to, and south of, the middle portion of the Cuyama River which is the long north branch of the shorter Santa Mafia River shown discharging into the ocean at about lat 35° in Fig. 1. 3 4 SETTE J. Res. Lepid, individuals (this is the type locality of macaria Edws.); the exact fre- quency of these types is not known; but there is a high percentage of silvered and unsilvered present ... In the Greenhorn Mountains, the segregation into a silvered population on the eastern side of the summit and an unsilvered one on the western side is decidedly apparent, though mixing occurs toward the south, where the populations unite”. The main purpose of this paper is to describe semi-quantitatively the variation of silvering within the Southern Zone of Intergradation. The silvery decoration of animals is far less common than the pigmen- tal coloration. While Hovanitz ( 1941 ) has found general qualitative correlations of the pigmental coloration of ^butterflies with environ- mental conditions, he finds the significance of silvering in callippe "quite incomprehensible" (Hovanitz, 1943, p. 420). Some peculiari- ties of the silvering data revealed by the semi -quantitative treatment in this paper suggests that studies of its physico-chemical nature and genetic control might be rewarding. An hypothesis is advanced as a point of departure for such studies. THE SAMPLES The material for this study consists of complete series of callippe that I collected during June 8 to 15, 1957 in company with Fred T. Thorne, and one sample from Bouquet Canyon consisting of part of a series taken in 1952 by Thorne and kindly given to me. In brok- ing the series Thorne exercised no conscious selection for color or pattern, though worn and damaged specimens were discarded. Parallel or convergent evolution among the Argynnids in some regions presents problems of species identification. In the Southern Zone of Intergradation this problem arises between calliiype and cor- onis Behr. The latter, as the subspecies hennei Gunder, flies with cal- lippe and resembles it so closely that confident identification requires either great familiarity with both species or known series from the region for comparison. L. P. Grey is in possession of both and kindly reviewed my field identifications. These had been made in consulta- tion with Thorne. At the time neither of us had previous acquain- tance with hennei, though we were both familiar with several other subspecies of coronis. Two specimens I had labelled callippe proved to be undoubted coronis. One labelled callippe came back from Grey with the notation, "Could this be coronis? (Don’t ask me!." Another labelled coronis bore Grey’s notation "I think it is a callippe!* The last two, when spread and compared with the now-available good series of hennei, appear to me to be undoubted coronis, though the under side wing surface (the only surface exposed to Grey’s examina- tion) is remarkably like the callippe from the same locality. These four speciments have been excluded from the samples. In the list that follows, the collections are grouped by localities. Each group contains mainly specimens taken at one location, to which have been added a few specimens from nearby locations. Those i(j):}-20, 1^62 VARIATION IN SILVERING 5 in the main group, I believe, are members of a single interbreeding colony; the others may consist of strays from the same colony in some instances, and in others, may be strays from some other colony. All localities are in Kern County except Sandberg’s and Bouquet Canyon which are in Los Angeles County. Distances from named places are given to the nearest mile (I.6I km) measured on a straight line on large-scale road maps. Directions, also determined from such maps, are referenced to true North and stared in Mariners’ abbreviations appropriate to the 32 -point compass. Thus southwest is siven as SW, southwest by west as SWxW and west southwest as WSW. These are 11.25° steps and allowing for measuring inaccuracy are correct to about the nearest 15°. Elevations were taken by aneroid altimeter graduated in 100-foot (30.5 m) units and probably accurate to ± 200 feet. Botanical names are according to Jepson (1923 - 1925), Abrams (1940, 1944 and 1951) and Abrams and Ferris (I960). No- menclature of the butterflies follows the list by McDunnough (1938) except for the genus {Speyeria), which, at the species and subspecies level follows the arrangement of Dos Passos and Grey (1947). A. East slope, Greenhorn Mountains. June 12, 1957, 18 males and 1 female taken 6 miles WSW of Kernville, elevation 4900 ft., along the road leading steeply from Isabella Reservoir to Greenhorn Mountain Park passing between Tittle and Rattlesnake Creeks, in association of Digger Pine (Pinus sabiniana Dough), Oak (Quercus sp), California Fremontia (Premontia cali- fornica Torr.), and chaparral broken by grassy areas, with Yerba Santa {Brio- dictyon so.) flowers as nectar attractant; included in addition is 1 male taken 8 miles WxS of Kernville, elevation 5500 ft., Transition Zone, in a small roadside opening in Yellow Pine (Pinus ponder osa Dough) and Incense Cedar {Libocedrus decurrens Torr.) forest. B. West slope, Greenhorn Mountains. June 12, 1957, 58 males and 6 females taken 27 miles NE of Bakersfield, elevation 3300 ft., near the foot of Eugene Grade, in Digger Pine, Blue Oak (Qurcus douglasii Hook & Arn.) and grass association, with violets (Viola sp.) abundant; included in addition are 3 males and 1 female taken 26 miles NNE of Bakersfield, elevation 3000 ft., in Blue Oak and grass association, with riparian flora along a small creek. C. Havilah. June 11, 1957, 13 males and 1 female taken 2 miles S of Havilah, elevation 3100 ft., Digger Pine, mixed oak and chaparral association with riparian flora along a small creek, mostly attracted to Yerb Santa flowers; in addition are 1 male from 1 mile north of Havilah, elevation 2800 ft., and another from 5 miles S of Havilah, elevation 3900 ft. D. Walker Basin. June 11, 1957, 30 males and 6 females taken 10 miles SxW of Havilah, at the south end of Walker Basin, elevation 3200 ft.. Blue Ok and Grass association with sagebrush and violets. Hoarhound (Marrubium vulgare L.) attracted a few individuals, but most were taken in the characteristic slow, fluttering flight displayed when on breeding grounds. E. Tehachapi. June 14, 1957, 12 males and 7 females taken 8 miles W of Tehachapi (the town), elevation 4700 ft, at the eastern end of Bear Valley in Blue Oak and grass association, with patches of sagebrush (Artemisia tri- dentata Nutt.), mostly acctracted to Wallflower (Erysimum sp.) flowers; in addition are 3 males from 7 miles W of Tehachapi, elevation 4600 ft., and 1 female from 3 miles SW of Tehachapi, elevation 4600 ft. 6 SETTE /. Res. tepid. I2 0‘ 119* 118® 36' 35* 3 4* FIGURE 1. Location of populations sampled (letters "A to H”) super- imposed on the distribution of Argynnis (Speyeria) callippe (shaded area) adapted from a portion of Hovanitz’ map (1943, Fig. 1) showing the distribu- tion of the "silvered spots” character by shading lines slanted upward to the left and the "unsilvered spots” character by shading lines slanted upward to the right. The Kern River (upper right) separates the Greenhorn Mountains to the north from the Piute Mountains to the south as it flows to a sink in the lower San Joaquin Valley (blank area at upper middle). At middle left is the Santa Maria River with the Cuyama River as its long north fork. At the lower right is the Santa Clara River. The crests of the higher mountains and ridges of the transverse series of ranges lie north of the Santa Clara River Drainage and south of the Cuyama River. F. Lebec, June 8, 1957, 24 males and 8 females; June 9, 1957, 26 males and 4 females; and June 15, 1957, 11 males and 15 males; total, 61 males and 27 females taken 1 mile S of Lebec, elevation 3600 ft., in oak and grass association with adjacent chaparral including sagebrush {A. tridentata) , Buck- wheat Brush {Eriogonum^ fasciculatum Benth.) and Yerba Santa; some were at flowers of the last, others in breeding-ground flight. (r;.-j-20, 1^62 VARIATION IN SILVERING 7 G. Sandberg^S (on old Ridge Route). June 9, 1957, 12 males and 3 females; June 15, 1957, 1 male and 2 females; total, 13 males and 5 females taken 9 miles ESE of Gorman in Digger Pine and sagebrush (not A. tridentata) associa- tion, with violets. Breeding ground behavior was predominant. H. Bouquet Canyon. June 11 and 12, 1951, 18 males collected by Fred T. Thorne at "upper Bouquet Canyon, 4 miles down from Pine Creek — Lake Hughes road” (estimated 16 miles W of Palmdale, estimated elevation 3400 ft.). HABITAT AND MESOSCALE DISTRIBUTION According to these records callippe in this region occupies ele- vations from 2800_ to 5500 feet, but occurs in good concentrations only between 3100 and 4900 feet, generally, the upper portion of the Up- per Sonoran Life Zone. Its habitat is marked characteristically by scat- tered Blue Oak, Digger Pine, and sparse chaparral with sagebrush {A. tridentata) often pesent. Usually there are broad areas or small patches of grass present. Practically all of the localities are in grazing land, but at the time of our trip it was being very lightly grazed. Violets were noted at only three of the locations. Where they were seen, the plants were abundant, mature, with fully-developed seed cap- sules and appeared about to dry up. Where violets were not seen, either they had dried up already, or the collecting was of members of the colony that had been drawn by nectar sources somewhat away from the site of the larval food plant. During the trip we collected at 31 locations within the geographic range and elevation limits of this butterfly. Of these, seven obviously were in, or close to, callip'be colonies; five locations yielding only one or several specimens, evidently were not as close; and the remaining 19 locations where we took none probably were distant from colonies, though this may not have been true for a few places where we sighted an Argynnid or two v/hich could have been either this species or coro- nis. These indications favor a population model consisting of well- separated, compact colonies from which individuals do not stray often or far. This pattern is more accentuated in this region than in most others within my collecting experience. Butterflies of 30 other species were taken at one or another of the seven locations where callippe were caught in good numbers. Plebeius acmon West, and Hew. and Hesperia lindseyi Holland were the most ubiquitous, occuring at most locations from the Greenhorn Mountains to Sandbergs. The former were taken in small numbers while the latr ter, in some locations were very abundant. In addition to these, the most common and prevalent in the Greenhorns and the Piutes were Euphydryas chalcedona Doubleday and Hewitt, Melitaea palla Boisdu- val, and Strymon saepl^m^ Boisduval; while through the Tehachapi to Sandberg’s they were Minois silvestris Edwards and Argynnis coronis Behr. If one of the specimens whose identification was questioned by 8 SETTE /. Res. Lepid. Grey is, as I believe, coronis, this species extended north at least to Havilah. Inasmuch as all coronis taken on this trip were fresh males, and in other regions within my collecting experience coronis flies some- what later than callippe, it is likely that a week or two later in the season coronis would have been found to be even more consistently a companion of callippe. RELATIVE ABUNDANCE The relative abundance in butterfly populations may be signifi- cant in studies of variation. From records extending through many years, Ford (1945) found that a rapid increase in abundance of the colonial butterflv Euphydryas aurina Rott was accompanied by "an ex- traordinary outburst of variation". When the increase ceased, the colony settled down to a comparatively uniform type, different from the one prevailing prior to the increase. Scientists differ as to the in- terpretation and significance of this phenomenon as a mechanism in evolution. Whatever the interpretation, if the generality of this phen- omenon is to be ascertained, records must be accumulated through the years for a number of butterfly populations. To this end, there are placed on record in Table 1, estimates of relative abundance of callippe in each of the seven colonies sampled in 1957. This estimate is in terms of catching rate and assumes that the number caught in unit time is proportional to the abundance of the population in the area of collecting. It is homologous in concept to the measure "catch per unit of effort” almost universally and suc- cessfully (sometimes "calibrated” to a absolute abundance by tag-and- recapture experiments) used as a basic statistic in studying the dy- namics of fish population fluctuations (Ricker 1940). For callippe I have computed the catching rate, R, according to the formula: R ^ Nc . H - k Nt where Nc is the number of callippe caught, H is the time, in hours, in the collecting period, Nt is the total number of specimens of all species caught (including callippe) during the collecting period, and k is a constant whose value depends on the methods and dexterity of the collector in caring for a specimen once it is in the net. By timing myself in the_caring, individually, for each of 150 specimens of a number of species from all families except Megathymidae under typ- ical field conditions, I found my k value to be 0.0155 hours (37.8 seconds ) . In repeated samplings of several species I have fo*und R to be sur- prisingly stable. For callippe, the three samplings at Lebec are an ex- ample of this. Although stable, R as computed by this formula should i(i):3’2o, I $62 VARIATION IN SILVERING 9 be regarded only as a first-order approximation, or index, of true rel- ative abundance, because it is affected by systematic error and several sources of variability, not all of them random. A more sophisticated iterative treatment theoretically should improve R, but it involves as- sumptions of uncertain merit awaiting test. TABLE 1. Relative abundance in terms of catching rate, R, in number per males and by the per cent of the males that were fresh. Catch- ing Per Per cent Locality, date (1957) and elevation rate cent of males (R) males fresh A. E. slope, Greenhorns, el. 4900 ft., June 12 ....... * 39 84 86 B. W. slope, Greenhorns, el. 330 ft., June 12 ........ **227 91 52 C. Havilah, el. 3100ft., June 11 16 93 15 D. Walker Basin, el. 3200 ft., June 11 * 44 83 87 E. Tehachapi, el. 4600 - 4700 ft., June 14 33 67 79 F. Lebec, el. 3600 ft., June 8 24 75 84 F. Lebec, el. 3600 ft., June 9 24 87 69 F. Lebec, el. 3600 ft., June 15 30 42 45 G. Sandberg’s, el, 4000 ft., June 9 17 80 75 In the meantime, of most concern is that R increasingly underes- timates abundance as the latter increases, presenting simultaneous catching opportunities more often and usually avail can be taken of only one of them at a time. Recognizing this, the values in Table 1 that probably were moderately underestimated owing to this factor are marked with an asterisk, and the one value that was grossly under- estimated is marked with a double asterisk. At the other end of the scale, when only one or two specimens are caught during a collecting period, R has large error owing to random variability of incidences of encounters. Avoiding this. Table 1 gives catching rates computed for only the first-listed location under each locality, except for Tehach- api where the times and catches of the first two locations were pooled for computing the catching rate. Another error source is that as kNt approaches H, a small error in k produces a large error in R. This approach was not close enough to be critical for the values given in Table 1, though the value for the west slope of the Greenhorn Mountains may have been moderately affected. Every collector can think of a number of other obvious sources of error or variability such as weather conditions, unusual concentrations at attractants, etc.. These obvious ones can be avoided by comparing only sets of samplings taken under reasonably similar conditions, as is true for the set in Table 1 except as later noted. Of course the catching rate reflects the relative abundance only at the stage of flight at which the collecting was done, and may yield values departing substantially from the inherent abundance. An esti- 10 SETTE /. Re$. tepid. mate of the stage of the flight is afforded by the relative numbers of fresh and worn individuals. In Table 1, recorded as *'fresh” were in- dividuals without readily perceptible random loss of scales or fringe. Damage reasonably attributable to accidental causes, such as notches, tears, breaks and rubbed streaks or patches was not considered. The data are given for males only, there not being enough females to yield reliable percentages. The two sexes were not combined for this sta- tistic because this would introduce extraneous variability owing to the tendency of females to emerge later than males. As evidence of this, only four of the 53 females listed in Table 2 were worn. More direct evidence is afforded by the Lebec samples. Those of June 8 and 9 were mostly males and mostly fresh. About a week later more than half of the males were worn and the females, all fresh, outnumbered the males. The sample from Havilah is anomalous in respect of percentage worn in relation to sex ratio. Although most of the males were worn, only one female was taken. Probably the greater ease of collecting males at the Yerba Santa flowers diverted us from the females which probably were widely scattered for ovipositing in the adjacent area which ap- peared to be the likely habitat for violets, though the plants were not observed, doubtless having dried up before our visit. At Sandberg’s the catching rate was substantially depressed bv a strong breeze sweep- ing the exposed slope and wafting many a disturbed butterfly out of stalking or pursuit range. Appraising the catching rates in the light of these qualifying fac- tors, the presence of some worn males at all localities indicates that the callippe flight was well developed throughout the region. But the high percentage of fresh males and the scarcity of females suggests that the flight had not yet reached its peak at most localities except Lebec where it probably was near peak stage on June 15 and Havilah, where the male flight had apparently passed its peak. Allowing for the differ- ences in stage of flight and the collecting difficulties at Sandberg’s, it appears that relative abundance was capable of attaining a height-of- flight index of 30 to 40, probably closer to the latter, at all localities except the location on the west slope of the Greenhorn Mountains where abundance was at least an order of magnitude higher, and at Walker Basin where it probably was substantially higher than 40 per hour, though well below that on the west slope of the Greenhorn Moun- tains. Although we did not trace out the geographic extent of a colony in any of the localities, the impression gained while collecting, was that these two colonies covered more extensive areas than the rest. It may be concluded, first, that the samples were drawn from near the middle of the flight period and hence probably represent nearlv the modal characteristics of the population in each colony; and second, the colonies varied quite widely in population size, perhaps through more than one order of magnitude. Further results may emerge if compar- isons with changed levels of abundance become possible in the future. 1^62 VARIATION IN SILVERING 11 VARIATION IN SILVER SCALING The silver scaling in callippe^ as in Argynnids generally, is confiined to certain pattern elements on the ventral wing surface. For the most part these are well defined spots in the apical and subapical area of the fore wing and all well defined spots in all areas of the hind wing. In addition, on heavily silvered individuals, silver scales may form ill defined streaks along the inner margin and in some of the interspaces between spots in the basal and discal areas of the hind wing. The pres- ent study is confined to the well defined spots of the hind wing. When unsilvered these spots are a light buff color, usually sharply bordered or outlined by brown scales. Some or all of the spots may be partly sil- vered. Then the silver scales and buff scales stand out in sharp con- trast to each other when specimens are held in the best relation to the light source and the eye to bring out the specular quality of the silver scales. A fully quantitative measure of the degree of silvering would be the ratio of silvered scales to all scales in the area potentially subject to silvering. To avoid spending a prohibitive amount of time count- ing scales, I have employed a subjective system of scoring based on general appearance. It would have been desirable, too, to utilize ex- clusively the pristine fresh individuals, but to avoid reducing the num- ber below levels needed for tests of significance, I have scored the slightly and moderately worn individuals along with the fresh ones, ex- cluding only those whose loss of scales caused serious doubt as to their correct score. The penultimate column of Table 2 shows the number excluded for this reason. Because the amount' of silvering sometimes differed appreciably in the basal and discal area, which will be called "disc” for brevity, from that in the submarginal area, called "margin”, the two portions of the wing were scored separately. Four grades of silvering were recognized: Grade 0 for complete lack of silver scales: Grade 1, when there were only flecks of silver well separated by buff scales; Grade 2, when some of the spots were unsilvered, others were partly silvered and still others were fully silvered, or any combination of two of these conditions; and Grade 3, when all spots were so fully silvered that there were no readily perceptible buff scales. The Disc Grade and the Margin Grade were treated as half scores and added together to arrive at the individual Score. Thus there are seven possible scores: 0 to 6, inclusive. Of the 238 individuals scored, 50 had half scores that differed from each other by one grade point. None differed by more than one. Of the 50 with different half scores, the Disc Grade was higher than the Margin Grade for 40 individuals and lower for 10. Curiously, for the two Greenhorn Mountain samples and the' Havilah sample, pooled, the ratio was 33:4, while for the Walker Basin, Tehachapl and Lebec sam- ples, also pooled, it was 7:6. The difference between the ratios 33:4 12 SETTE /, Re$. tepid. and 7:6 is statistically significant (p=0 02) desoite the small numbers in the latter. It is possible that the tendency for the marginal row of spots to be less silvered than the disc spots is linked to the general lack of silvering in the three northernmost localities. TABLE 2. Frequency distribution of silvering scores Sex and sample Score 0 1 ' 2 3 4 5 6 Mean Scored Total Males : A. E. slope. Greenhorns 5 3 1 1—2 1 3.16 — - 19 B. W. slope, Greenhorns 17 8 7 64 8 7 2.42 7 64 C. Havilah 2 1 — 2 — 2 4 3.73 4 15 D. Walker Basin 1 2 1 — 1 2 19 5.08 4 30 E. Tehachapi 1 — — - i— _ — 13 5.40 — 15 F. Lebec 1 1 1 1 1 1 39 5.54 15 60 G. Sandberg’s — — — — — . — 10 6.00 3 13 Total 27 15 10 11 6 15 99 4.16 33 216 Females : A. E. slope, Greenhorns — 1- — ■ — — — -2.00 — ■ 1 B. W. slope. Greenhorns — — 2 3 - — — — 2.60 5 C. Havilah l l D. Walker Basin _ „ _ 11 1 3 5.00 — 6 E. Tehachapi _ _ l 1 6 5.01 — 8 F. Lebec 1 _ _ _ 25 5.77 1 27 G. Sandberg’s - — — — . — — — 3 6.00 — 5 Total 1 — 4 5 1 i 39 5.12 2 53 The mean score for males is 4.16 and for females is 5.12 (Table 2), suggesting that females tend toward more silvering than males. But four-fifths of the females were taken at Walker Basin, Tehachapi and Lebec where silvering in both sexes is nearly complete. If we pool these three sampls and cHss together the individuals with scores 0 to 4 inclusive, and similarly those with scores 5 and 6, in order to have high enough numbers for statistical test, we find no significant dif- ference between the sexes (P = 0.9). Accordingly, further analysis deals with both sexes combined, as shown graphicallv in Fig. 2. The arrangement of samples in Fig. 2 is in ascending order of their means. This also arranges them from north to south except for the reversal of samples A and B. Where samples are near the same par- allel of Latitude and have appreciable meridianal displacement, as among A, B and C as one group and F, G, and H as another, their means also ascend from west to east. The slopes of the lines connecting the means in Fig. 2 should not be taken as representing gradients of silvering in the sense of reflecting unit increase of silvering per unit distance. The panels are not spaced in proportion to distance, but merelv to accommodate the height of the bars. The means, themselves, are defective because the scores do not represent even gradations in amount of silvering. Further, because of t§€a VARIATION IN SILVERING 13 wesr SLO PE. GREEN - HORN MTS. 35* 37' EAST SLOPE, GREEN - HORN MTS. 35® 43 ' HAVILAH 35® 31' WALKER BASIN 3 5* 23' TEHACHAP! 35®08' LEBEC 34*50* SA NDBERG'S 34® 45' BOUQ UET CANYON 34® 35' FIGURE 2. The degree of silvering of the several samples, shown as histo- grams of the number of individuals according to grade (see text); and the mean grades, shown as small circles connected with heavy straight lines. NUMBER OF INDIVIDUALS 14 SETTE J. Res. Lepid. the prevalence of low numbers at the middle scores, the distributions depart widely from the bell-shaped normal probability curve and the par? metric test commonly applied to means would not yield reliable probabilities even if the means were not otherwise defective. The means were computed and connecting lines were supplied in the graph only to give a general perspective of the amount and direction of dif- ferences between localities. For a more careful assessment, the significance of the differences between samples, as to the numbers of individuals falling into the sev- eral silvering classes, can be made by the chi -square test which re- quires neither a normal distribution nor uniform class size. It does require a pooling of the original score classes to provide a minimum of five as the expected number on a class. To meet this requirement, the data have been grouped into two classes: (1) individuals with Scores 0 - 4, which can be described as "unsilvered” and partly silvered”; and ( 2 ) those with Scores 5 and 6, which can be described as "silvered or nearly so”. With this grouping the approximate probability, P. of the difference between any two samples in respect of the ratio of "un- silvered and partly silvered” to "silvered or nearly so” occurring by chance can be found by entering the table of chi-square distribution on one degree of freedom with the value of chi scfuare computed from the 2x2 contingency table according to the method given by Fisher (1948, p. 85). For greater precision the Yates correction for contin- uity was incorporated and P was found by graphical interpolation be- tween tabular values. The results are given in Table 3 as a matrix in which the chi-square values betyeen all combinations of pairs of sam- ples are shown in the upper right portion and the corresponding values of P in the lower left portion. TABLE 3. Chi-square values (upper right) and probability values (lower left) between pairs of samples. Locality Sample and sample B A C D E F G W. Greenhorns .. B 2.37 2.97 >10 >10 >10 >10 E. Greenhorns .. A 0.16 0.02 4.63 1.10 >10 7.14 Havilah C 0.08 0.10 1.24 1.74 0.83 1.78 Walker Basin D < .001 .03 .19 .01 .84 2.32 Tehachapi E <.001 .02 .18 .10 .64 1.36 Lebec F <.001 <.001 .34 .34 .40 .37 Sandberg’s — G <.001 .008 .02 .14 .19 .10 The three northern samples. A, B, and C form a group within which there is no statistically significant difference at the five per cent prob- ability level (P = 0.05), and the four southern samples, D to G, in- clusive, form another such group. But these groups are not discrete from each other. Sample C from Havilah differs no more significantly from D, E, and F than from A and B, indeed, somewhat less. The sum of chi squares among A, B and C is 5.36; df are 3; and P is 0.14. The i(i):3-20, 1962 VARIATION IN SILVERING 15 comparable values among C, D, E, and F are: Chi-square, 5.34; df, 6; P, 0.48. The indeterminate position of Havilah relative to the two groups, most probably is due to the small number in the sample from there. With such a small sample it would require a very strong con- trast in silvering ratio to prove statistical significance. Until more ma- terial is available it remains uncertain whether the population at Havi- lah resembles more closely those to the North or those to the South in respect of silvering, or whether it is truly intermediate as suggsted by Fig. 2. Likewise, the sample from the east slope of the Greenhorn Moun- tains is small and does not test significantly different from that of the west slope. But for the Greenhorn Mountains there are two sources of additional data. Thorne and I scored both his catch and mine while in the Greenhorns and this record includes about twice as many indivi- duals as given in Table 2. We used three categories: ' unsilvered”, *'part silvered”, and “silvered”, corresponding somewhat inexactly to the Scores 0 - 1, 2 - 4 and 5 - 6, respectively, of Table 2. Hovanitz (1934) reported the number of individuals in a sample from the east side of the Greenhorn Mountains, elevation 5500 feet, and from the west side at Cedar Creek, elevation 5000 feet, according to three cate- gories: “not silvered or very slightly so”, “intermediate” and “well - or fairly-well silvered”. Assuming these categories correspond approx- imately to those of Thorne-Sette, the data may be pooled as follows: East side: Unsilvered Intermediate Silvered Total Thorne - Sette . ......6 .15 17 38 Hovanitz ....0 ...... ..2 14 16 Total West side: ...........6 17 ......... 31 ...54 Thorne - Sette ... 72 ...24 32 128 Hovanitz ....10 8 4 22 Total .............. .........82 ...... ..........32 36 150 With more data it is possible to use a 2 x 3 contingency table for a more discriminating test. This yields a probability far less than one in a thousand that the west side and east side samples could have been drawn from a population containing the same proportions of unsil- vered, partly, silvered, and silvered members. This does not prove that there is no mixing between the populations of the west and east slopes; it only indicates the extreme unlikelihood of enough inter- change to form a thorough mixture. But when it is considered, in addition, that the ridge of the mountain is clothed with transition forest within which no colonies of callippe were found, it seems clear that the interchange, if any, must be slight and probably by way of mixing with the more silvered populations to the South at different rates on the two sides of the mountain range as suggested by Hovan- itz (1943, p. 411). 16 SETTE /. Res. tepid. Turning southward, past Havilah, there is at Walker Basin a small proportion of unsilvered and partly silvered individuals in the pop- ulation. The low proportion continues with only slight further dim- inution in going from Walker Basin in the Piute Mountains, through the Tehachapi Mountains to Lebec. This uniformity suggests a rela- tively brisk interchange between colonies in this stretch of mountains, probably coupled with strong environmental selection against unsil- vered individuals. Then at Sandberg’s, only a short distance beyond Lebec, all indiv- iduals are silvered, This is true also for the specimens from Bouquet Canyon (not included in the scoring table because they were only a part of a series). But my sample from Sandberg’s is too small to assuredly include unsilvered or partly silvered individuals if they con- stitute only a small percentage of the population. There is historical evidence, in Gunder’s (1930) list of butterflies of Los Angeles Coun- ty, of the occurance of unsilvered callippe at several localities in the transverse ranges south and southeast of Lebec. Under ”Araynnis ma- caria lamina Wri. THE UNSILVERED MACARIA FRITILLARY, a transition form.” Gunder recorded one male collected by Comstock "June 10, 1922, Ridge Route;” one male collected by Friday "June 19, 1929, Pine Canyon;” and two males and one female collected by Gun- der "June 25, 1921, Bouquet Canyon.” "Ridge Route” probably is identical to, or at least near, the locality I give as "Sandberg’s”; "Bouquet Canyon” may be identical to, or near the locality of the Thorne col- lection here listed under the same place name; Pine Valley lies about half- way between the two. The last is definitely on the desert side of the transverse ranges, while the other two localities are near the crest divid- ing the drainage to the desert from the drainage to the Pacific. The Thorne collection is definitely from the Pacific drainage side. It is proba- bly that this crest marks the southern limit of the unsilvered or partly silvered form. Reviewing, there is a mixture of unsilvered, partly silvered and fully silvered callippe extending through the Southern Zone of Inter- gradation, from the Greenhorn Mountains in the North, southward through the Piute and Tehachapi ranges, to the crest of the series of ranges extending transversely across California at about latitude 34° 40’ N. Most of the change from the predominantly unsilvered con- dition in the north to the predominantly silvered condition in the South, take place in about the northernmost one-fourth of the Zone. Through the remaining three-fourths, beginning at about the mid- length of the Piute Mountains, where a high degree of silvering al- ready has been reached, the increase in silvering is very gradual. Al- though it is still perceptibly incomplete at Lebec near the Southern end of the Tehachapi, unsilvered and partly silvered become so rare beyond there that one should not expect to encounter evidence of the unsilvered condition without drawing large samples from the popu- (i):}-20, ip62 VARIATION IN SILVERING 17 lation. It is reasonable to conclude that this is the southern limit of the tendency toward lack of silvering in the ventral hind-wing spots of callippe. To express this finding in Fig. 1, the area shaded with lines slanted upward to the right should be extended to include localities E, F, G and doubtfully H. When so extended, the southern limit of lack of silvering would more nearly agree with the southern limits of "uniformly colored red-brown upper surface” and "lack of brown pig- ment between sdo<-s on the underside of the hind wings” as mapped by Hovanitz (1943, Figs. 2 and 4). DISCUSSION Although this description adds some details to the distributional pattern of silvering in callippe, it raises more questions than it an- swers. Some are concerned with geographic limits. How much far- ther north does silvering extend? Does it taper off gradually from the western Greenhorn Mountains northward as does the opposite condition from Walker Basin southward, or does it end abruptly a short distance north of the sampled Greenhorn Mountain localities? Does the more silvered population of the eastern slope of the Green- horn Mountains end as in a cul-de-sac co-terminal with the limits of the Kern River drainage, or does it extend northward as a tongue along the higher Sierra Nevada toward silvered populations such as those near Downieville and Gold Lake? More perplexing questions concern the processes governing silver- ing. The distribution of individuals among the seven score classes re- veal a paucity of intermediates in the middle of the range (histo- grams A and C in Fig. 2 ) . Could such a frequency distribution arise from simple Mendelian inheritance, or is a more complex mechanism required? The existance of intermediate deo:rees of silvering suggests sim- ilarity with the example of simple Mendelian inheritance described by Ford (1945, p. 173) for the Currant Moth, Abraxas grossulariata L., which, in its commonest form, has wings with a white bick.Erround. In its less comomn form, lutea Cockerell, the white is replaced by a deep yellow. When these two forms are interbred the heterozygotes have a pale yellow background. Mating an individual homozygous for white with one homozygous for deep vellov/, produces offspring inter- mediate between the two parents. There is no dominance. Taking this as a model for silvering in callippe, and using Ford’s system for notation we may use the symbols and for represen- ting the unsilvered ( Score 0 ) , the partlv silvered ( Scores 1 - 5 ) , and the completely silvered (Score 6) phenotypes, respectively. For such a model, in a population containing equal numbers of the two homo- zygotes (C^C^ and OC^), we should find : C^C^: 1:2:1. None of our samples have equal numbers of homozygotes, but if we pool samples A, B and the ratio is 24 : 51 : 18. This 18 SETTE /. Res. Lepid. differs from the ratio 1 : 2 : 1 by no more than one would expect from random sampling variability in about half of the samples drawn (P =z 0.44). We may conclude that there is no evidence against the hypothesis that silvering is controlled by a single pair of alleles of the phenotypes scored 1 to 5 may be lumped as the heterozygous com- ponent of the population. However, in the Currant Moth, equal doses for white and for deep yellow produced a color about half way between the two homo- zygotes. In callippe there are very few individuals about half silvered. Individuals scored 3 and 4 comprise less than one-third of the cate- gory with scores 1 to 5. Perhaps some more complex genetic system may account for this peculiar distribution within the partly silvered segment of the population. But I am attracted by the idea that some chemico - physical process, such as crystallization may be involved. These tend to be triggered by very slight differences in conditions, and once triggered, go to completion. In this connection it is also interesting that in areas of potential silvering, the single scales, as viewed with a hand lense, appear either totally buff or totally silver, never in between. In pursuing this idea, I have found only a little information on the silvery decorations of the Lepidoptera in the literature that bear on its chemical nature. Mayer (1897) makes direct reference to its nature in the Argynnids, saying ”... Dimmock ('83) has shown that the silvery white and milk-white colorations are due to optical effects produced by reflected light. In the silvery white scales, however, such as the under surface of the hind wings of Argynnis, there must be a reflecting surface toward the observor, for both silvery and milk- white colors appear simple milk-white by reflected light.” According to Fox (1953, p. 289) uric acid, derived from chrysalid metabolism, deposited in the wings of the adult butterfly; guanine and uric acid contribute opaque whiteness and glistening silvery aspects. Ford (1945), although discussing extensively the physical and chemical nature of butterfly coloration, is silent on the nature 'of silvering. Taylor, (1925) points out that guanine (also spelled guanin) deposited in crystalline form imparts the silvery appearance prevalently displayed by pelagic fishes, while bottom fishes with prevalently white under- surfaces have their "subdermal tissues heavily charged with amorphous guanin, which is chalky white.” He describes guanine crystals, after they have been processed into pearl essence for making artificial pearls, as ". . . very thin light blades, floating in a liquid . . . [which] show their maximum luster when they are oriented parallel to each other . . . [when] pointing promiscuously in all directions, the effect will be a metallic or dull pearly luster.” While one would prefer evidence derived directly by analysis of the substance as it occurs in callippe, it is not far fetched to suppose that this substance is guanine, sometimes deposited as crystals disposed t(i):3-20, i$6a VARIATION IN SILVERING 19 with appropriate orientation to give the specular effect that we call silvery, and sometimes in amorphous form adding whiteness to the pigmental brown to produce the light buff of the unsilvered spots. On this supposition it is necessary that the "silvering gene” control only the conditions within the pupal scale-sac fluid so as to precipitate guanine amorphously for buff scales, or alternatively, as crystals for the silver scales. Presumably, in homozygotes the balance of influences is well to one side of the critical point so that the deposite in all scales is always amorphous, or always crystalline. In the heterozygotes, too, it may take place on the "all or none” basis, but by single scales. Under the equal and opposite influence of the two alleles the equilib- rium may be very unstable and readily tipped to one side or the other of the critical point. This could be mediated by very slight differences in environmental conditions of the particular microclimate of a pupal individual at the precise time of color deposition in the scale sacs. This would tend to affect nearly all of the scales of an individual alike, swinging most of them toward one side of the critical point more often than nearly equal numbers to each side. Small differences between members of the colony as to pupation site, date and time of day of deposition of salts in the wing sacs, or even the rate of drying after eclosion, might provide the variety of conditions necessary to produce the observed peculiar frequency distribution of the supposed hetero- zygous phenotypes. This set of ideas is not advanced as an explanation, but as an hypothesis inviting test by those having interest and competence in the fields of experimental breeding or biochemistry, or both. It is a pleasure to acknowledge the stimulus to undertake this analysis and the material help received from William Hovanitz; the generosity of Fred Thorne in giving specimens and sharing his uncom- mon collecting serendipity; and the kindness of Paul Grey in review- ing my identification of many hundreds of specimens of this and other Argynnids. LITERATURE CITED ABRAMS, LEROY. 1940-1951. Illustrated Flora of the Pacific States, vols. 1-3, 4 vols. Stanford. ABRAMS, LEROY and R. S. FERRIS. I960. Illustrated Flora of the Pacific States, vol. 4, Stanford. DOS PASSOS, C. F. AND L. P. GREY. 1947. Systematic Catalogue of Speyeria with designation of types and fixation of type localities. Amer. Mus. Novitates, no. 1370: 1-30. JEPSON, W. L. 1923-25. A Manual of the Flowering Plants of California. Berkeley. . FISHER, R. a. 1948. Statistical Methods for Research Workers. Hafner, N. Y. 10th Ed. GUNDER, J. D. 1930. Butterflies of Los Angeles County, California. Bull. Sou. Calif. Acad. Sci. 29 (2) ; 39-95. HOVANITZ, WILLIAM. 1941. Parallel ecogenotypical color variation in butterflies. Ecology 22: 259-284. 20 SETTE J. Res, Lepid. 1943. Geographical variation and racial structure of Argynnis callippe in California. Amer. Nat. 77:400-425. MAYER, ALFRED G. 1897. On the color and color-patterns of moths and butterflies. Bull. Mus. Comp. Uool. Cambridge 30 (4) : 169-256. RICKER, W. E. 1940. Relation of "catch per unit effort” to abundance and rate of exploitation. /. Fishery Res. Board Canada, Toronto. TAYLOR, HARDEN F. 1925. Pearl essence: its history, chemistry, and technology. Appendix to: Rept. U.S. Commissioner Fisheries for 1923. Bur. Fish. Doc. 989. Washington, D.C. 1(1) :21.42, 1962 Journal of Research on the Lepidoptera 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Copyright 1962 THE EFFECT OF VARIOUS FOOD PLANTS ON SURVIVAL AND GROWTH RATE OF PIERIS^ WILLIAM HOVANITZ AND VINCENT C. S. CHANG California Arboretum Foundation, Inc., Arcadia^, Los Angeles State College, Los Angeles and University of California, Riverside The close relationship between most phytophagous insects and their host plants is a general phenomenon known by most entomologists. This relationship is so outstanding that field naturalists can often locate colonies of any particular insect most readily by first searching out the host plants, as they often are larger and more obvious to the eye than the insects themselves. What is not so well known is ( 1 ) the extent to which it is possible for any particular species or race to survive on various food plants, and (2) the causes of the attraction to, and survival of, any race of insect on a particular plant. It is the purpose of this paper to indicate something of the range of ability to survive of two insects, Pieris rapae and Pieris protodice, (the "cabbage or mustard "butterflies) on various cruciferous plants. In other papers, it is intended to go into the various aspects of the development of food plant preferences in insects, including its causes, evolution, and relationship to origin of phylogenetic groupings. THE MATERIAL AND METHOD The experimental work to be reported on here and in subsequent papers has been carried out in Arcadia, Southern California. All species of the genus Pieris (Lepidoptera: Pieridae) found in southern Cali- fornia have been utili2ed for the study. The geographical ranges and specific food plants of the species existing in southern California are being described elsewhere (Hovanitz, 1962) and the general distribu- tion of the species of North America in another paper (Hovanitz, 1962). Five species of Pieris exist in this area, P. rapae (a European immigrant), P. protodice, P. beckeri, P. sisymbrii and P. napi. Each of these occupies a specific habitat and is restricted to a different food plant.^ 1 Aided by a grant from the National Science Foundation, Washington, D.C. 2The authors wish to express their great appreciation to Dr. W. S. Stewart, Director, Los Angeles State and County Arboretum for great cooperation in making this work possible. 3Some taxonomists consider Pieris occidentalis to be a distinct species separate from Pieris protodice, rather than a geographical race, ecological race or local phenotypic form. If this were the case, there would be six species in southern California. 21 22 HOVANITZ AND CHANG /. Re$. Lepid. Pieris rapae is the easiest species to breed under greenhouse condi- tions and has been used for most experiments. It aslo has a wider range of satisfactory food plants than the others. As a European immigrant, it remains restricted largely to plants of European origin and has never adapted itself under natural conditions to any American native. Contrariwise, Pieris protodice, a native species, has never adapted itself to any European plant grown in cultivation, but has assumed an "economic” role only in utilization of European plants (mustards) growing in semi-wild habitats. The other Pieris are restricted entirely to one or a few native species of plants. The food plant tests were made in several series at different times of the year and with different species of plants. Each series was run for the purpose of comparing the influence of several plants (usually five) on growth rate and mortality. For each series, a group of twenty larvae, obtained as follows, were grown on each plant. Female Pieris rapae were allowed to oviposit on a kale plant in the greenhouse. For the first series, the eggs were removed from the kale leaf and placed on leaves of the plant being tested in a petri dish. On all other tests, the eggs .were allowed to hatch on kale, and to eat on it for one or two days before removal to the tested plants. In order to minimize the effects of variable environmental conditions, all tests of a series were grown at the same time and place. For example, eggs laid on September 11, 1959, were used for the first test. On September 14, 1959, the larvae were measured for the first time, and then successively thereafter each day until pupation. The dates of larval (or pupal) deaths were noted. The later tests were similar except that (1) the origin of the parents was different, ( 2 ) one or two day old larvae were used instead of eggs, ( 3 ) the time of year and ( 4 ) place of breeding was different. Data on origin of the material for each test are indicated in Table 1. TABLE 1 : Source of larvae used in these experiments. Series Species Food Source Locaiity ( coilected Date Generations of experient from wild Western 1 Pieris rapae Cabbage Orange Sept.11,1959 1 County 2 Oct. 9,1959 2 3 Nov. 25, 1959 4 4 Feb. 20, 1960 6 5 Feb. 20, 1960 6 6 Aug. 2, I960 11 7 Sept. 19, I960 12 8 Pieris protodice Mustard Laguna Beach Aug. 27, 1959 1 9 II San Fernando Valley July 18, 1960 1 10 Cleome Owens Valley July 21,1960 1 i(t):2i-4S, 1962 EFFECT OF FOOD PLANTS 23 SERIES NO. 1 Series No. 1 was conducted for the purpose of comparing the effect of the following plants on growth rate, size and mortality: black mustard (Brassica nigra), garden nasturtium (Tropaeoum majus) , bladder pod (Isomeris arborea) and watercress {Nasturtium officinale) ( Table 2 ) . The females used for the source of larvae were obtained from an open cabbage field isolated from any other cruciferous food plant source and over which the adults were swarming in huge num- bers. The probability that they came from any other source is infinitesimal. TABLE 2. Viability and growth rate of Pieris rapae grown on different kinds of food plants ( Series 1 ) . Plant Number Number and percent of larvae died Days from egg to pupation Range of variation in larval growth Nasturtium 20 16 30% 20 3 Kale 20 2 10% 12 2 Mustard 20 8 40% 13 3 Watercress 20 18 90% 17 1 Nasturtium 20 16 80% 18 5 Isomeris 20 10 50% 20 3 The mortality (Table 2. Series 1) was highest on watercress (90%), next highest on garden nasturtium (80%), then Isomeris (50%), mustard (40%) and least on kale (10%). The rate of development as determined by the first larva to pupate, was most rapid on kale (12 days), least rapid on Isomeris (20 days), and increasingly more rapid on nasturtium (18 days) , watercress (17 days) and mustard (13 days). The range in variation of larval growth was greatest for nasmrtium, since five days elapsed between the first larva to pupate and the last. The range as indicated in the table should be interpreted with the fact in mind that only a fraction of the original twenty larvae remained at the time of pupation. In other words, only two larvae pupated out of twenty that were fed on watercress. Had more survived, the range between the dates of pupation might have been greater. The growth rates of the larvae on the various plants were obtained by measuring ten larvae of each group each day and obtaining the average length. These were then plotted and a curve drawn for each. These curves are shown in Figure 1. Measurements of the larvae were started on the third day after hatching as they were so small prior to that time. However, great ■^It should be noted that the watercress used in these experiments was obtained by purchase in a food market while all other plants were grown in the laboratory. It has occassionally been found that despite repeated washings, insecticide residues may be present upon such commercial vegetables and that these may cause death of larvae feeding upon them. 24 HOVANITZ AND CHANG I. Res. Lepid. differences already were apparent between the larvae grown on the various plants. Those on kale were 4.7mm long as compared with those on Isomeris which were only 1.5mm long. Those on nasturtium were 1.8mm, those on watercress were 1.9mm and those on mustard were 3.9mm. The kale-group larvae were over three times the size of the Isomeris group larvae. Increase in size of the larvae on subsequent days was fastest v/hen the larvae were grown on kale, second fastest on mustard, and about equally as poor on each of the other three plants. The largest sized larva was that grown on Isomeris (24mm) as com- pared with that on kale (22.2mm) but after eight more days of larval life. Those on watercress and nasturtium pupated at 20mm and those on mustard at 17.9mm. It appears that the greatest proportional size difference in the larval groups occurred during the first three days, but that most of the loss in size was later recovered by a longer growth period. SERIES NO. 2 The tests of Series No. 2 were carried out in order to correct what were thought to be defects in the experiments of Series No. 1. These were two in number. The first was the high mortality in the eggs when they were removed from kale and placed on the other plants. This problem was averted by letting the larvae hatch on the kale plant before moving to the tested plants. The second presumed defect was the possibility of increased mortality of the very young larvae on the new plants, due solely to physical characteristics, toughness, etc. This was partially averted by allowing the larvae two days of feeding on kale before transference to the tested plants. One change in food plant was made in this test: radish obtained from a food market was employed in place of watercress. Another inevitable change was the climatic condition under which the tests were conducted. They were changed slightly by the fact that the tests were carried out starting October 9, 1959, instead of September 11, 1959. TABLE 3. Viability and growth rate of Pieris rapae grown on different kinds of food plants (Series 2). Days Range of from variation egg to in No. and % No. and % Total No. pupa- larval Plant No. of larvae died of pupae died and % died tion growth Kale 20 2 10% 0 0 2 10% 11 2 Mustard 20 2 10% 3 15% 5 25% 13 4 Radish 20 8 40% 5 25% 13 65% 13 5 Nasturtium 20 8 40% 2 10% 10 50% 14 5 Isomeris 20 12 60% 3 15% 15 75% 15 3 The mortality (Table 3) in this series was generallly not so high PIER IS RAPAE Growth rates of larvae of Pieris rapae grown on various plants ( Series No. 1 ) . 26 HOVANITZ AND CHANG /. Res. Lepid. as in the series just preceding, though that on kale was the same at 10%, those on mustard dropped from 40% to 10%, on nasturtium dropped from 80% to 40% and on Isomeris increased from 50% to 60%. Watercress was not tested this time but instead radish^ was used, with a resulting mortality of 40%. The better survival of these larvae is probably correlated with the fact that they were all grown during the first two days on kale. Generally, the time (minimum days) to pupation in this series has been decreased (Fig. 2), The time on kale has been decreased one day, on mustard it remains unchanged, and on nastrutium and Isomeris it has been reduced five days. These reductions are probably correlated with the more successful survival of the young larvae on kale for the first two days, as compared with Series No. 1. SERIES NO. 3 The tests of Series No. 3 are identical with those of Series No. 2 with the exception that they were started November 25, 1959 instead of October 9, 1959; in addition, the larv.ae by this time were four generations removed from the wild and inbred. The larval mortality for all plants other than nasturtium was in- creased considerably in Series No. 3 compared with Series No. 2; 10% to 30% for kale, 10% to 35% for mustard, 40% to 80% for radish and 60% to 75% for Isomeris (Table 4). The data are com- parable for the percentage of pupae died. TABLE 4. Viability and growth rate of Pieris rapae grown on different kinds of food plants ( Series 3 ) . Days Range of from variation Plant No. No. and % of larvae died No. and % of pupae died Total No. and % died egg to pupa- tion in larval growth Kale 20 6 30% 0 0 6 30% 15 2 Mustard 20 7 35% 0 0 7 35% 16 5 Radish 20 16 80% 1 5% 17 85% 19 4 Nasturtium 20 7 35% 1 5% 8 40% 16 6 Isomeris 20 15 75% 2 10% 17 85% 18 8 The length of larval life has also been increased considerably in Series No. 3 tests, and thus also the range of variation in larval growth. The reasons for these differences between Series No. 2 and No. 3 are probably two fold. First, the later time in the season has meant that the conditions of growth were somewhat different, primarily in the temperature. The larvae were grown in a greenhouse in which the temperature was not well-regulated. Therefore, the temperature of the October ^The radish used in this particular experiment, but not in others reported on by us, was obtained in a food market. Refer to footnote (4) for possible latent effects. (mm) 3VAMV1 JO H10N31 (i):2i-42, 1962 EFFECT OF FOOD PLANTS 27 tests (Series No. 2) was higher than for the November tests (Series No. 3). This would profoundly influence the results insofar as the absolute growth rates are concerned. On the other hand, it did not affect the relative growth rates on the various food plants. These are still in the same descending order: kale, mustard, radish, nasturtium and Isomeris. 0 ■ I 1 ■ ! I I ■ ■ U— I I U 2x 3 4 5m 6 r 8 9x )0 ItxK !2 13 14 15 DAYS AFTER HATCHING FROM EGG UNTIL PUPATION NO OBSERVATION MX = NO OBSERVATION ON KALE 2. Growth rates of larvae of Fieris rapae grown on various plants (Series No. 2). The differential increase in size of the larvae on the various food plants is better shown in Series No. 3 (Fig. 3) than in Series No. 2 ( Fig. 2 ) . This is because the slower development rate served to exaggerate the differences between each strain. The larvae grown on kale grew larger at a faster rate than the larvae on any other plant, followed by mustard, nasturtium, Isomeris and radish. In Series No. 2, the order of the latter two were reversed. It is interesting to note that the ultimate size of the larvae which grew slowest was seldom equal to the size of those larvae which grew fastest. In fact, there is almost a direct relationship between these events. A second reason for the difference between the mortality and growth data of these two series may be of genetic significance. The extent of inbreeding in the laboratory population by the fourth genera- tion would have been sufiicient to (a) increase the mortality, (b) slow 20 0 1_ P/ERIS RAPAE HOVANITZ AND CHANG J. Res. Lepid. 28 O (u/m) 3VAi/Vl JO H19N31 DAYS AFTER HATCHING FROM EGG UNTIL PUPATION (x=NO OBSERVATION) (i):2i-42, 19^2 EFFECT OF FOOD PLANTS 29 development rate, (c) decrease size, (d) magnify the difficulties of feeding on abnormal plants by the simple means of decreasing overall vigor by the accumulation of allelic recessive lethal genes in the laboratory strain which had not yet had a chance to be eliminated by the selective effects of homozygosity. SERIES NO. 4 AND 5 Series No. 4 was a test to compare various garden vegetables for their usefulness to Pieris rapae. These vegetables were mustard, kale, kohlrabi, broccoli and brussels sprouts. Series No. 5 was a similar test for various other plants. These were mustard (as before), nasturtium, radish, turnip and Isomeris. The tests were made in two series on February 20, 1962, using larvae following six generations of inbreeding, but of the same strain as used heretofore. In both of these tests, the larval mortality was higher on comparable plants used in previous tests. This again was probably due to the effects of inbreeding. TABLE 5. Viability and growth rate of Pieris rapae grown on different kinds of food plants (Series 4). Days Range of from variation Plant No. No. and % of larvae died No. and % of pupae died Total No. and % died egg to pupa- tion m larval growth Mustard 20 9 45% 0 0 9 45% 15 9 Kale 20 5 25% 2 10% 7 35% 15 10 Kohl Rabi 20 9 45% 0 0 9 45% 14 9 Broccoli 20 9 45% 0 0 9 45% 14 9 Brussels Sprouts 20 10 50% 0 0 10 50% 14 9 In Series No. 4 (Table 5), the data show that the larvae on kale had the lowest mortality rate (25%) while all the other plant tests were about the same (45-50%). Despite this, however, larvae from kohlrabi, broccoli and brussels sprouts reached pupation ahead of those on kale or mustard. TABLE 6. Viability and growth rate of Pieris rapae grown on different kinds of food plants ( Series 5 ) . Plant No. No. and % of larvae died No. and % of pupae died Total No. and % died Days from egg to pupa- tion Range of variation in larval growth Mustard 20 9 45% 0 0 9 45% 15 9 Nasturtium 20 19 95% 0 0 19 95% 16 Radish 20 14 70% 2 10% 16 80% 15 13 Turnip 20 13 65% 0 0 13 65% 15 4 Isomeris 20 12 60% 0 0 12 60% 18 14 30 HOVANITZ AND CHANG /. Res. tepid. In Series No. 5 (Table 6), the data show that the lowest mortality again was on mustard but with considerably higher mortality on the other plants: for example, 95% on nasturtium, 70% on radish and 60-65% on turnip and Isomeris. The length of larval life was longest on Isomeris (18 days) and about the same (15-16 days) on other plants. The range of larval growth period varied greater with radish and Isomeris than with any of the other. 4 5 6 7 ex 9 10 II 12 13 14 I5x . 16 DArS AFTER HATCHING FROM EGG UNTIL PUPATION tx^NO OBSERVATION) 4. Growth rates of larvae of Pieris rapae grown on various plants (Series No. 4). The rate of increase in size (Fig. 4) was more rapid with broccoli, kohlrabi and brussels sprouts than with kale and mustard. It is believed that this may be correlated with a more turgid or water-content condi- tion of these plants as compared with the drier or stiffer condition of the kale and mustard. For some reason, in Series No. 5, also, the increase in size of the larvae was not up to expectations on the mustard since the larvae on radish, turnip and nasturtium all exceeded those on mustard in final size ( Fig. 5 ) . SERIES NO. 6 Series No. 6 was tested on August 2, I960 with the use of larvae eleven generations from the wild. One new plant, {Cleome lutea) , was tested in this series, since this plant was found to be primary food plant for Pieris protodice in Owens Valley, California. Mortality data PiERiS RAPAE i(i) 121-42, 1^62 EFFECT OF FOOD PLANTS 31 o o 00 in o in cJ q o CVJ (u/m) BVAyVl JO H10N31 Growth rates of larvae of Fieris rapae grown on various plants ( Series No. 5 ) . 32 HOVANITZ AND CHANG /. Res. Lepid. for kale , mustard, nasturtium and Isomeris, (Table 7) compared well with previous data obtained on these plants (Series 1, 2, 3, 4 and 5). Cleo7ne was only slightly less effective in maintenance of larval life, the mortality rate being 30% as compared with 20% for kale and mustard. The relative lengths of larval life and the range of variation in larval growth also compares favorably with previous data. The higher mortality as seen in Series 3, 4, 5 and 6 has here been reduced to nearly the level of Series No. 1 . TABLE 7. Viability and growth rate of Pieris rapae grown on different kinds of food plants (Series 6). Plant No. No. and % of larvae died Days from egg to pupation Range of variation in larval growth Kale 20 4 20% 12 3 Mustard 20 4 20% 13 3 Nasturtium 20 10 50% 13 3 Cleome 20 6 30% 14 5 Isomeris 20 12 60% 18 6 The differences between growth curves for each of the plants tested has been greatly increased in this test as compared with earlier tests. The factors responsible for this are unknown but may be due to a reduced adaptability caused by increased homozygosis (Fig. 6). The larvae on kale were larger than those of any other plant at all times during their growth period. Pupation occurred the twelfth day. The larvae grown on nasturtium were almost the same size as those grown on mustard, whereas in Series No. 1 (Fig. 1) the larvae on mustard grew faster and larger. The larvae on Cleome were fourth in size on any day, while those on Isomeris were fifth. SERIES NO. 7 This series was conducted in order to place the plant Thelypodium affine into position relative to the other standard test plants in the laboratory. Thely podium affine is a native cruciferous plant from the desert regions which is not known to be a food plant of any Pieris species, though another Thely podium species is the food plant of Pieris dsymhrii. TABLE 8. Viability and growth rate of Pieris rapae grown on different kinds of food plants ( Series 7 ) . Plant No. No. and % of larvae died Days from egg to pupation Range of variation in larval growth Kale 20 5 25% 16 3 Mustard 20 6 30% 17 4 T. affine 20 14 70% 22 5 Isomeris 20 12 60% 18 6 Nasturtium 29 8 40% 17 6 PIER IS RAPAE ■{1)121-42, 1962 EFFECT OF FOOD PLANTS 33 \ o o CSJ fluuij 3VAHW1 JO HI9N31 DAYS AFTER HATCHING FROM EGG UNTIL PUPATION 34 HOVANITZ AND CHANG /. Res, Lepii. The tests of Series No. 7 (Table 8) were carried out on September 19, I960, using larvae after twelve generations of laboratory inbreeding and growing on kale. The larvae grew relatively well on mustard, kale and nasturtium. The effectiveness of Isomeris was as previously tested. Thely podium affine was the poorest of all these plants as far as effectively providing the larva with stimulation to feed. The mortality of the larvae on T, affine was 10% compared with 60% on Isomeris, 40% on nasturtium, 30% on mustard and 25% on kale. The length of larva life was increased from 16 days on kale to 22 days on Thelypodium. The final size of the larvae was reduced from 22mm long on kale to only 18mm long on the Thelypodium larvae which survived. (Fig. 7). GENERAL OBSERVATIONS ON PIERIS RAPAE TESTS The seven series of tests on Pieris rapae larvae have indicated several significant points: ( 1 ) Larvae from strains grown previously on cabbage or kale have a lower mortality rate, have a faster development rate, and a greater size when grown on kale than when grown on any other plant tested. When similar larvae are grown on mustard, these factors are decreased slightly. The order of decrease for plants other than these two is in the order approximately as follows: nasturtium, Cleome, radish, Isomeris, and T, affine. Plants of the cabbage group: kohlrabi, broccoli and brussels sprouts, are about the same as cabbage or kale. Turnip is about the same as radish. ( 2 ) Higher mortality in the tests is generally correlated with a slower development rate, a delayed time to pupation and smaller size of the larva and pupa. There are some exceptions to this rule; for example, the final size of the larva may be as great when the growth rate was slow on one plant as when it was fast on another. This has been indicated, however, only for plants of the cabbage group. (3) Mortality rates in the various test series have been variable. These variations in mortality have been thought not only to be due to the genetic effects of inbreeding but also partly to the changed environ- mental conditions encountered while the tests were conducted, such as lower temperatures, etc. of the changing seasons. Lower temperatures slow the developing rate, and increase the relative humidity. Increased humidity permits the larvae to be attacked by bacterial or viral diseases to a much greater degree. SERIES NO. 8 Tests of Series No. 8, 9 and 10 were conducted with larvae of Pieris proto dice rather than with Pieris rapae. The first series with this species (No. 8) was conducted on August 27, 1959, utilizing larvae one generation from the wild. The fwu/} 3VAHV1 JO H10N31 DAYS AFTER HATCHING FROM EGG UNTIL PUPATION ( NO OBSER V ATION 36 HOVANITZ AND CHANG /. Res. Lepid. females of the previous generation were obtained in a field of black mustard, not near any other source of food plant, and in which eggs and larvae were found on that plant. This is good' evidence that the natural plant utilized here as larval food was mustard. In this series, the food plants tested as larval food were mustard, kale, radish, Isomeris and nasturtium. The mortality of larvae of Pieris protodice on these plants increased in the order just stated, namely, 10% on mustard, 20% on kale, 55% on radish, 55% on Isomeris and 95% on nasturtium (Table 9). Deaths in the pupa stage on kale, radish and Isomeris materially increased the total mortality before adult emergence. This was not increased on mustard. TABLE 9. Viability and growth rate of Pieris protodice grown on different kinds of food plants (Series 8). Plant No. No. and % of larvae died No. and % of pupae died Total No. and % died' Days from egg to pupa- tion Range of variation in larval growth Mustard 20 2 10% 0 0 2 10% 10 2 Kale 20 4 20% 2 10% 6 30% 11 2 Radish 20 11 55% 4 20% 15 75% 13 3 Isomeris 20 11 55% 3 15% 14 70% 16 3 Nasturtium 20 19 95% ■— — - 19 95% 21 The length of larval life was greatly different on the various food plants, ranging from ten days on mustard, eleven days on kale, thirteen days on radish, sixteen days on Isomeris to twenty-one days on nastur- tium. This wide variation in length of larval life is greatly different from anything encountered with Pieris rapae. It would indicate a fundamental difference in food plant adaptability between these two species, Pieris rapae being much more versatile in its requirements than Pieris protodice. The size of the larvae at various days after hatching (Fig. 8) indicates that the mortality data are in close correlation with the size increments of the larvae with respect to food plants involved. Larvae on mustard and kale increased to 24mm in length, those on radish, Isomeris and nasturtium to 20-2 1mm each. SERIES NO. 9 Discovery that Pieris protodice utilizes Cleome lutea as a natural food plant in the Owens Valley prompted a series of tests involving this plant. Both Series No. 9 and Series No. 10 were carried on for the purpose of testing Cleome in comparison with other plants used as standards. Series No. 9 utilized Pieris protodice larvae bred from adults obtained from mustard fields in the San Fernando Valley while Series No. 10 utilized Pieris protodice larvae obtained from Cleome pf£ms pporoDicE i(i);2i-42, i$62 EFFECT OF FOOD PLANTS 37 mrs AFTER HATCHING FROM EGG UNTIL PUPATION (NO observation = x j 38 HOVANITZ AND CHANG /. Res. Lepid. plants in the Owens Valley (Big Pine). There is no chance that the San Fernando population had access to Cleome but there is a chance that the Owens Valley population did have limited access to mustard. TABLE 10. Viability and growth rate of Pieris proto dice grown on different kinds of food plants (Series 9). Plant No. No. and % of larvae died Days from egg to pupation Range of variation in larval growth Mustard 20 6 30% 14 3 Cleome 20 7 35% 14 4 Nasturtium 20 18 90% 17 2 Isomeris 20 8 40% 17 5 The Series No. 9 larvae (Table 10) showed highest mortality on nasturtium (95%), just as in Series No. 8, with the lowest mortality on mustard (30%), next highest on Cleome (35%) and then Isomeris (40%). This corresponds well with comparable data in Series No. 8. The growth increment curves for this series are somewhat com- parable with the preceding series except for the one surviving larva on nasturtium which increased to a size greater than that attained by larvae on any other plant. The differences between the growth curves, however, are not as distinct and clear as in Series No. 8, perhaps because of temperature differences between the two. Cleome is almost as satisfactory as mustard as a food plant for these larvae. ( Fig. 9 ) • SERIES NO. 10 This series was run almost contemporaneously with the preceding and involved larvae from the Owens Valley strain grown on Cleome. The mortality (Table 11) was lowest on mustard (20%), second TABLE 11. Viability and growth rate of Pieris protodice grown on different kinds of food plants (Series 10). Plant No. No. and % of larvae died Days from egg to pupation Range of variation in larval growth Mustard 20 4 20% 11 3 Cleome 20 6 30% 12 5 Isomeris 20 7 35% 13 7 lowest on Cleome (30%) and then Isomeris (35%). There was a greater range of larval growth rates on this group of larvae on Isomeris than in any other group as well as the lowest mortality on this plant in any of the larvae tested. The growth increment curves for these larvae indicate that the larvae do not show much difference in their reaction to mustard and Cleome but that mustard is still slightly better, and Isomeris is always poorer. Final size of the larvae, however, was larger for the Cleome (23 mm) than for the mustard (22 mm) or the Isomeris (18 mm)'. (Fig. 10). LEN6TH OF LARVAE (mm) LENGTH OF LARVAE (> i(i):2\-42, 1962 EFFECT OF FOOD PLANTS 39 DAYS AFTER HATCHING FROM EGG UNTIL PUPATION 9. Growth rates of larvae of 'Pieru protodice grown on various plants (Series No. 9). PAYS AFTER HATCHING FROM EGG UNTIL PUPATION 10. Growth rates of larvae of Pieris protodice grown on various plants (Series No. 10), 40 HOVANITZ AND CHANG /. Res. Lepid. GENERAL OBSERVATIONS ON PIERIS PROTODICE TESTS The three series of tests on Pieris protodice have indicated several significant points: (1) Larvae from strains obtained from mustard or Cleome populations had equally good survival value on mustard plants. The following plants gave decreasing survival value to the larvae: (Series 8) kale, radish, Isomeris and nasturtium. (Series 9) Cleome, nasturtiuna, Isomeris. (Series 10) Cleome, Isomeris. (2) Higher mortality in the tests is generally correlated with a slower development rate, a delayed time to pupation and smaller size of larvae and pupae. DISCUSSION The tests described in the preceding pages have indicated some significant information regarding the effect of various food plants on survival in two species of Pieris. The ability to survive on different food plants is not an all or none relationship, but rather is a relation- ship based upon relative ability to survive. In the most extreme case, all larvae may die rather than eat a completely unacceptable plant. On the other hand, plants may be placed in an order of desirability with regard to the ability of larvae to survive, or to grow satisfactorily, on plants which they may accept at least reluctantly. The tests described in this report are based upon using three criteria to indicate the desirability of a plant to the larvae. The first of these is the ability to survive, or the reverse which is the mortality, expressed as a percentage of the total larvae which died during the experiment. The second criterion is the daily size increment during the life of the larva. This actually tests two factors: (1) the physiological usefulness of the food to the larvae and (2) the total amount of food eaten and uilized, which may be a direct result of the stimulatory activity of the plant toward the larvae. The third criterion is the range in variation of larval growth. The significance of this criterion is in more doubt than the other two, but the range indicated would in some degree show the genetic variability present in the larval population toward physio- logical adaptation to the particular plant. The results just described show that some plants are much more suitable for larval food to certain larvae than are other plants. The reasons for this greater suitability are not made clear. In those cases where higher mortality, slower growth rate, size increments and final size are correlated, the cause of the poor growth could be lack of nutriment in the plant, or lack of larval ingestion of the plant. In other cases where larval mortality may be high, as with the broccoli, Kohl- rabi and Brussels sprouts tests (Table 5) hut growth rate rapid and size large ( Fig, 4 ) , there are probably two factors involved. The rapid increase in size could be due to a high attraction to a very nutritious plant, but secondary conditions surrounding the plant may be conducive I (i):2i -42, 1962 EFFECT OF FOOD PLANTS 41 to a higher mortality, namely, high water content in the plant itself, or higii humidity surrounding tne plant whicn would increase the susceptability of the larvae to virus and bacterial intections. Toughness of leaves of a plant cannot be ruled out as a factor involved in larval speed of growth. Some leaves are too tough for the larvae to feed upon easily. Leaves of the same plant vary in tnis regard; succulent leaves at the siioot apex are usually much more desirable than older, mature leaves. This may be caused by toughness, or it may be caused by the greater perception of attractive substances permeating from younger leaves. At any rate, the relative uniformity of the tests in the difterent series is good evidence for the conclusion that this factor has not been of major significance. The greater ability of tieris rapae as compared with Pieris protodice to survive on a variety of common cruciferous plants is interesting in view of the fact that both succeed well on black mustard. Pieris rapae survives better on kale over mustard and P. protodice survives better on mustard over kale. However, on most other cruciferous plants tested, there was a wide disparity between their preferences. The selections made by the strains of larvae shown by these experiments complements perfectly the known differences in geographical and ecological distribu- tion of the two species concerned, and indicates that there is little natural competition between the species for food plants. This shows also why Pieris rapae is always associated with European cruciferous plants, weeds or truck crops, while Pieris protodice is generally associated with other cruciferous plants. Takata (1957, 1959, 1961) has published a number of papers which bear upon this problem. None of these deals precisely with the effect of various food plants on survival and growth rate of Pieris, but rather deals only with selection preferences by the larvae and the adult. Further considerations of the relation of this work of Takata on Pieris rapae crucivora to the present general program will therefore be made in a complementary paper entitled "Three Factors affecting larval choice of food plant” to be published in this journal. A visit to our project by W. H. Dowdeswell of England during the course of these experiments was followed by a similar study of more limited extent on Pieris napi L. in England. Dowdeswell found that two populations of this Pierid were feeding on different cruciferous plants. By switching the food plants in laboratory experiments on test- ing the larvae, he found as we had, that there was a higher mortaity and decreased growth rate when the larvae were fed on plants not the normal food plant for that population (Dowdeswell and Willcox, 1961). SUMMARY 1. Tests were made of the ability to survive of larvae of two species of Lepidoptera, Pieris rapae and Pieris protodice on various food plants. 42 HOVANITZ AND CHANG /. Res. tepid. 2. Three criteria were tested: (a) mortality, (b) growth rate and size increment, (c) range in variation of larval growth. 3. Seven series of tests involving Pieris rapae were conducted showing the results summarized under the heading "General Observa- tions on Pieris rapae Tests.” 4. Three series of tests involving Pieris protodice were conducted showing the results summarized under the heading "General Observa- tions on Pieris protodice Tests.” 5. Pieris rapae has significantly different requirements for food plants than Pieris protodice, even though the tested larvae were both from mustard-bred strains. 6. The native adaptability of Pieris rapae to survive on various plants is better than that of Pieris protodice. 7. The causes of differences in survival value of the larvae on different food plants are believed to be several, but the main one is the total amount of food eaten and utilized. LITERATURE CITED DOWDESWELL, W. H. AND H. N. A. WILLCOX. 1961. Influence of the food plant on growth rate and pre-imaginal mortality in the green-veined white butterfly Pieris napi (L). Entomologist 94:2-8. HOVANITZ, W. 1962. The ecological and geographical distribution of Pieris species in Southern California. J. Res. Lepidoptera 1 : in press. HOVANITZ, W. 1962. The distribution of the species of the genus Pieris in North America. J. Res. Lepidoptera 1 ( 1 ) : this issue. TAKATA, N. 1957. Studies on the host preference of cabbage butterflies {Pieris rapae L.) III. Mechanism of food preference of cabbage butterfly larvae. Jap. J. Ecol. 7 (3) :117-119. TAKATA, N. 1959. Studies on the host preference of common cabbage butterfly, Pieris rapae crucivora Boisduval. IV. Tendencies of food prefer- ence in relation to total quantity of the host plants cultivation. Zoological Magazine (Japan) 68(5) : 187-192. TAKATA, N. 1959. Studies on the host preference of common cabbage butterfly, Pieris rapae crucivora Boisduval. VI. Change in the food preference of larvae when reared successively by the definite food plant for several generation. (Preliminary report) Jap. J. Ecol. 9:224-227. TAKATA, N. 1961. Studies on the host preference of common cabbage butterfly, Pieris rapae crucivora Boisduval. XI. Continued studies on the oviposition preference of aduir butterflies. Jap. J. Ecol. 11 (3) :124-133. TAKATA, N. 1961. Studies on the host plant preference of the common cabbage butterfly, Pieris rapae crucivora Boisduval. XII. Successive rearing of the cabbage butterfly larva with certain host plants and its effect on the oviposition preference of the adult. Jap. J. Ecol. 11(4): 147-154. Journal of Research on the Lepidopvera l(l):43-49, 1962 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Copyright 1962 GENERAL CHARACTERISTICS OF THE MOVEMENTS OF VANESSA CARDUI (L.) J. W. TILDEN San Jose State College, San Jose, Calif. It is a matter of common knowledge that the populations of many species of butterflies participate in large and extensive mass movements. While several workers, notably C. B. Williams of England, have done extensive work in this field, relatively few investigators have made it their principal subject of research. Though much pertinent information has been assembled and certain tentative con- clusions may be reached, it cannot be said that the movements of butterflies are as well understood as certain other phases of the study of these insects. Most movements of butterflies, while usually referred to as migra- tions, would not meet the criteria set up by students of vertebrates for defining migrations. The migrations of birds, for example, involve cyclic movements. There is a going out and a return. Few butterflies other than the Monarch [Danaus plexippus (L.)], habitually possess such a rhythmic pattern of movement. Even with the Monarch, the individuals that return are usually (probably always) not the same individuals that initiated the outward movement, but are the offspring of these original individuals. Many movements of butterflies are relatively local wanderings, and may be in one direction in one local area, but in a different direction in another area a relatively few miles away. For example, in the fall of I960, Tilden noted Libytheana hachmanii (Kirt.) (the Snout Butter- fly) moving southwesterly in large numbers in the vicinity of Con- tinental, Pima Co., Ariz., on Sept. 14. The next day, Sept. 15, equally large numbers of this species were observed moving up-canyon ( north- ward) in Sycamore Canyon, Santa Cruz Co., about thirty -five miles south of Continental by air line, though considerably further by road. The very conspicuous movements of certain species have been regarded as depending on population pressures, leading to a mass exodus. These may be considered as emigrations from the populated area, and as immigrations into surrounding area. Such movements frequently appear to be in one direction only. Unless the insects were to find suitable conditions at the far end of the movement, establish- 43 44 TILDEN /. Res. Lepid. merit would not result. Such movements have been thought to account in part for the finding of strays far beyond the normal ranges of the species. Movements of an altitudinal nature have been observed. Nymphalis calif ornica (Bdv.) has been seen ascending the Sierra Nevada of California in great numbers. Such movements may result in finding sufficient larval food plants for a subsequent brood. In I960, Tilden found this species in outbreak numbers of adults at Manzanita Lake, Lassen Volcanic National Park, Calif., on June 10-17. Numerous larvae in various stages of development were found on Ceanothus spp. Adults could be observed many miles below the Park, working their way up the mountain. It may be significant to note that this species passes the winter as an adult. It seems plausible that these altitudinal move- ments of adults are the usual way for adults to reach the higher eleva- tions, where they act as parents of a late high elevational brood. This might be a fruitful field of investigation. The often spectacular "migrations” of Vanessa cardui (L.) present an interesting condition somewhat intermediate between true migra- tions and other types of movements. Return movements seem not to be clearly documented in regard to this species. However, reproduction has been observed repeatedly along the route, and the offspring take up the movement when they emerge as adults. These mass movements of Vanessa cardui thus proceed in periodic waves, alternating with relative lulls between, and the whole process has much more continuity than is characteristic of the movements of most other butterfly species. Much of the basic work on the movements of Vaness-a cardui has been done by C. H. Abbott. Others, including Sudgen, Woodbury and Gillette in Utah, have made contributions. Among the various titles that form the literature on the movements of this species in western United States, some are short observational notes only. Others present a considerable amount of detailed study. Abbott has attempted to understand the nature of these movements and has produced several penetrating papers on the subject. The last great movement was in 1958. Abbott traces three or more generations extending up to or beyond the Central Valley of California. Other generations carried the waves through the Bay Region of California on into Oregon. The generations involved in the northern portions of the movement become less clear. The first generation seems to have originated in northern Mexico, the second in the Imperial Valley of California, western Arizona and the eastern Mohave Desert of California, the third in western Mohave Desert, and the fourth in the Central and Coastal Valleys of California, and to the Bay Region. Abbott (1959) has presented material that includes the observations of a number of workers, and has coordinated this material. In addition, the follov/ing observations are here presented for the first time. j(i):4}-49, 1962 MOVEMENTS OF VANESSA 45 Tilden found many dead adults in western Arizona (Mohave County) and in the vicinity of Needles, California, as well as large numbers of moving adults, March 30-31, 1958. From Barstow, San Bernardino Co., west to Mohave, Kern Co., great numbers of adults were flying northwesterly on March 31. Later that day adults were seen moving through Tehachapi Pass, Kern Co., and spilling into the San Joaquin Valley east of Bakersfield, Kern Co. On April 20 to May 9, eggs and larvae were abundant in the Salinas Valley, Monterey Co. On May 30 - June 1, Tilden found large numbers of adults along the Redwood Highway (Route 101) between San Rafael, Marin County, and Cloverdale, Sonoma County, moving northwesterly. Dead adults were seen as far west as Booneville, Mendocino County, where but a few living adults remained. This may mark the western edge of the movement. In the vicinity of The Geysers, Sonoma County, adults were moving northwesterly in large numbers in overcast weather (fog) on June 1, before noon. Larvae were abundant on Amsinckia (Fiddleneck) and on the introduced Spanish Thistle ( Carduus ) , in various stages of development. Two interesting reports were from Oregon. Ray Albright reported large numbers of Vanessa cardui moving in a north-north-easterly direction on May 18-19, 1958, while on May 20-21, equally large numbers were observed moving in exactly the opposite direction. He attributes this reversal of direction to the insects meeting a cold front, since a storm was in progress some distance beyond the place toward which the insects were noted to be flying. The other report was from David Huntzinger at Crater Lake National Park. He noted that Vanessa cardui was abundant there in 1958, though scarce or absent in 1957. On April 1, 1958, Tilden observed what is apparently a new observation for the United States. Large numbers of Vanessa cardui were flying north-westerly in the rain. Between Vidal, San Bernardino County, and Desert Center, Riverside County, the rain was light to moderate, but rather cold. Between Indio and Whitewater Canyon, later the same day, the flight still continued in heavy to very heavy rainfall. This seems to indicate that adults may continue to fly, once on the wing, even in heavy precipitation. It would seem a priori that initiation of a flight during rain might be less likely. From observations so far recorded it is possible to generalize on some of the characteristics of the movements of Vanessa cardui. Some of these generalizations have been set forth by Abbott and others, while some are suggested here for the first time. 1. The adults tend to fly into the wind. This may be a key point in attempting to explain the prevailing direction of the movements. During the time of year when Vanessa cardui is moving, the prevailing winds in California are north-westerly, and the movements of the 46 TILDEN /, Res. Lepid. butterflies are also in a north-westerly direction, almost or quite directly into the prevailing winds. Sudgen, in 1937, noted a northerly direction of flight in Utah. Sudgen, Woodbury and Gillette, in 1947, noted a north-easterly direction of flight. Observations to be made in the future might very well take special note of the wind direction as well as the direction of the insects’ flight, to find how complete this correlation between wind direction and flight direction may be. 2. The adults move at a rather equal velocity in relation to one another, resulting in a fairly constant spacing of the individuals. There have been some observations made where this did not seem to be the case, but for the most part, this generality seems to hold. 3. The insects tend to fly over obstacles rather than around them. At times this appears to the observer quite amusing, even ridiculous. If a tree, building or parked car obstructs the route, the adults fly over the top of the obstruction, when in some instances it would appear that a veering to the right or left would be more easily accomplished. The obstruction, whether it be low or tall, is barely cleared. On March 30, 1958, Tilden observed them piling directly over a small building near Barstow, San Bernardino County, the only building for miles in an open desert. 4. The adults tend to fly at rather even densities. If a line is marked off between tv^o points, fifty or one hundred feet apart, and the insects crossing this line per minute are counted, the results from several one minute counts are surprisingly similar. If then, the observer drives his car several miles, sets up another station and makes another series of counts, the results will again be similar, and frequently quite close to those of the previous station. But in widely separated areas, the count may be very different. That is, while the density on the Mohave Desert, for instance, may be quite constant, counts made in San Diego County at another time might indicate a different density. Further study may shed light on the possibility that density of flight is related to density of population. 5. Vanessa cardui adults will fly in overcast, at least after the movement is under way. This has been observed between Victorville and Needles, San Bernardino County, March 31, 1958, and near The Geysers, Sonoma County, May 31 and June 1, 1958. 6. They have been observed once at least in the United States (vide supra) to fly in large numbers during moderate to very heavy rainfall. 7. The migrating individuals are all of one species. Other species of butterflies do not seem to become involved in the mass movements of Vanessa cardtii. 8. The individuals in the flights are of both sexes. This was noted independently by sampling in 1958, by both Tilden and O. E. Sette. The sex ratio of the insects when engaged in a flight is very nearly the same as that oi Vanessa cardui at any other time, 0.5 or one half i(i):43-49> 1962 MOVEMENTS OF VANESSA 47 of each sex. Tilden took a sample of fifty adults, and found twenty- seven males, twenty-three females. The numbers taken by Sette are not recorded. This phase of the investigation deserves further study and this is planned for the future. 9. The females from the above sample were not greatly gravid. None had eggs ready to oviposit. This suggests that the females may drop out of the flight when ready to oviposit. If this be true, it might explain the tendency to reproduce along the flight routes. 10. The total movement proceeds in waves, or phases, which so far appear to be based on broods, a conclusion which seems to be well supported by the work of Abbott. In the light of current findings, it is possible to reconstruct, at least to some degree, the nature of the populations of Vanessa cardui during and prior to the flights. 1. The first (initiating) brood apparently originates to the south, in northern Mexico or southwestern Arizona. 2. Findings from those movements that have been studied indicate that large flights can occur only in years when there has been sufficient rainfall on the deserts for a large growth of vegetation, enough to support such populations of Vanessa cardui. 3. The food plants in the desert areas, as noted by many observers, are principally Boraginaceae, especially Cryptantha, but also Amsinckia. Malvaceae have also^ been reported. Thistles and other composites seem to be of little importance at this time. 4. Upon emergence of the adults, each succeeding brood begins to fly into the wind, which in the deserts of California at this season (early spring) results in a north-westerly direction of flight. 5. These adults reproduce at some point along the line of flight. Time and place of reproduction presumably depend on the condition of the insects and the presence of suitable vegetation. Information on how far the females fly before reproducing, and the precise reason for selecting a certain place for reproduction, would add much to our knowledge. 6. Four to five broods, perhaps more, succeed one another before the force of the movement is spent. 7. The northern extent of the movement is at least to the San Francisco Bay Region of California. In some years, such as 1958, the effect extends into northern California, and apparently further, to Oregon and perhaps even Washington. Information as to the relation- ship between the Vanessa cardui in California and those in Oregon and Washington during such outbreak years is badly needed. Williams and Abbott have raised the question as to whether Vanessa cardui is found north of the Imperial Valley, Imperial County, Calif., in any but outbreak years. There seems to be good evidence that it is. An examination of pinned specimens from many parts of California shows specimens taken in almost every year for many years 48 TILDEN /. Res. Lepid. back. As a Professor at San Jose State College, Tilden has examined hundreds of student collections over a period of fourteen years, and has also examined other specimens dating back to 1920. Some speci- mens of Vanessa cardui appear in student collections every semester. It is evident that this species is a normal component of the butterfly fauna in many parts of California. It is probable, however, that this local endemic population has little or nothing to do with the mass movements which from time to time are superimposed on it. Two other items seem worth mentioning. Firstly, moderate to large populations of Vanessa cardui may be found late in the fall at high elevations in the mountains of the western United States. Specifi- cally, Tilden found hundreds of adults at Hannagan’s Meadows, White Mountains, Arizona, September 12, I960. A fair population was found at Barton Flats, San Bernardino County, Calif., in September, 1957. Similar populations have been reported from the Sierra Nevada of California and elsewhere. The nature of these populations is not known, but it would seem that they are independent of the populations concerned with the great flights that occur in certain years. Secondly, the mass flights of Vanessa cardui are of a very irregular nature. They are not regularly cyclic and do not occur at predictable intervals. However, some success in predicting these flights has been possible through knowledge of rainfall on the desert. Several workers foresaw the 1958 outbreak in the gathering populations of the south- west deserts. An interesting and provocative suggestion of how rainfall may be necessary for mass movements of this species, was observed by Tilden September 11, I960, at Yucca, Mohave County, Arizona. Late summer rains were plentiful and vegetation was well developed. Numerous freshly emerged adults of Vanessa cardui were present, and reproduc- tion was in progress, with larvae of various stages on Cryptantha. It seemed that the basis for a movement was all prepared. However, the winter rains did not materialize, and on March 26, 1961, the area around Yucca was dry and no Vanessa cardui were to be found. If this abortive season is contrasted with the abundant rainfall of 1958 over the entire area, it points to a dependence of these populations on ample winter rainfall in the areas that are far enough south to act as reservoirs for populations that can engage in a mass flight. Several questions are raised : ( 1 ) Why do these insects move at all, rather than remaining where they are? Possibly food may become ^scarce, but populations exist that do not move out in such a systematic manner under lowered food conditions. (2) Why is the flight directed? Why do not these insects move randomly in any direction open to them? The tendency to fly into the wind may be a clue here. Or have such movements been repeated so many times over the history of the species, that the tendency to fly in one direction has been selected for? (3) What causes the adults to reproduce here and there along the (i):43-49< ^9^2 MOVEMENTS OF VANESSA 49 route? Why do they not merely fly on until exhausted? Present information may suggest some tentative answers, but such answers, while plausible, are not yet supported by suflicient data. The author would add that this attempt at a synthesis of our knowledge concerning the movements of Vanessa cardui in western United States is prompted more by interest on his own part than by a desire to inform others. The purpose of this paper is to bring into focus our ignorance of the real basis of this interesting phenomenon, with a hope that future work will help to clarify some of the little known facets of the problem. LITERATURE CITED ABBOTT, C. H. 1941. The 1941 migration of the painted lady butterfly, {Vanessa cardui) in southern California. Bull. Ecol. Soc. Amer., 22:13. — — — 1946. Mapping the 1945 migration of Vanessa cardui in southern California. Bull. Ecol. Soc. Amer. 27 :49. — — 1950. Twenty-five years of migration of the painted lady butterfly, Vanessa cardui, in southern California. Pan-Pac. Ent., 26:616-172. — - 1951. A quantitative study of the migrations of the painted lady butterfly, Vanessa cardui L. Ecol. 32:155-171. — — 1959. The 1958 migration of the Painted Lady Butterfly, Vanessa cardui (Linnaeus), in California. Pan-Pac. Ent., 35:83-84. ESSIG, O. E. 1926. A butterfly migration. Pan-Pac. Ent., 2:211-212. SMITH, RAY F., & E. GORTON LINSLEY. 1945. Migration of Vanessa cardui (Linn.). Pan-Pac. Ent., 21:109. SUGDEN, JOHN W. 1937. Notes on the migrational flights of Vanessa cardui in Utah. Pan-Pac. Ent., 13:109-110. SUGDEN, JOHN W., ANGUS M. WOODBURY & CLYDE GILLETTE. 1947. Notes on the migration of the painted lady butterfly in 1945. Pan-Pac. Ent., 23:79-83. WILLIAMS, C. B. 1958. Insect Migration, xiii -f235 pp., 11 plates, 22 photos, 49 maps & diagrams. Collins, London. WOODBURY, ANGUS M., JOHN W. SUGDEN & CLYDE GILLETTE, 1941. Notes on migrations of the painted lady butterfly in 1941. Pan-Pac. Ent., 18:165-176. Journal of Research on the Lepidoptera 1(1) :51-61, 1962 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Copyright 1962 THREE FACTORS AFFECTING LARVAL CHOICE OF FOOD PLANT’ WILLIAM HOVANITZ AND VINCENT C. S. CHANG" California Arboretum Foundation, Inc., Arcadia, Los Angeles State College, Los Angeles and University of California, Riverside Prior to the initiation of comprehensive tests of the attraction by larvae of Pieris rapae of various food plants, a number of prelim- inary trials were made to test the effect of external factors which might influence the correctness of these tests. Of the factors which were thought most important, the following were selected for trial: ( 1 ) Influence of direction of air movement ( "wind”) . (2) Influence of age (size) of larvae. (3) Influence of food previously eaten by larvae, or strain of larvae. EFFECT OF WIND MOVEMENT An experiment devised to test this factor was set up as follows: A nursery flat about 24 inches square was filled with vermiculite. Pots of plants, each about the same size, were placed in the vermiculite spaced about as shown in Figure 1. The table under the flat was marked A, B, C and D. The direction of the wind movement was at side B. In the tests, larvae were placed in the central area, shown by a small circle and allowed to go to the plants, or to leave the flat. Ten different larvae were used for each experiment, and these were used a total of ten times each. Thus, each experiment involved 100 trials. The wind direction relative to- the position of the plants was changed by turning the flat by 90° after each 100 trials, as shown by the illustration (Fig. 1). The first position tested was with mustard toward the dirction from which the breeze came ( side B ) . The results of the preferential selection by the larvae are shown on Table 1. Kale was selected by 19%- of the larvae, mustards by 56%, radish 1 Aided by a grant from the National Science Foundation, Washington, D.C. 2The authors wish to express their great appreciation to Dr. W. S. Stewart, Director, Los Angeles State and County Arboretum, for great cooperation in making this work possible. 51 52 HOVANITZ AND CHANG /. Res. Lepid. TABLE 1. The effect of wind movement on the choice of food plants by larvae of Pieris rapae Kale Mustard Radish Nasturtium Isomeris None Total no. of trials A 19 B56 C18 D3 D 1 3 100 90° D20 A 20 B52 C4 C2 2 100 90° C43 D25 A 26 B2 BO 4 100 90° B53 C 15 D13 A8 A5 6 100 Ten 14.1 mm to 15.7 mm kale-bred larvae of Pieris rapae were used per test with each test ten times at each position. The parents of the larva were collected at western Orange County in a cabbage field. by 18%, nasturtium by 3% and Isomeris by 1%. Larvae which left the flat without selection amounted to the additional 3%. With the flat turned 90” so that the radish was now in position B ( in the direction of the wind ) , the selection results were very different. Kale remained relatively unchanged at 20%, mustard dropped from (i);5i~6i, 1962 THREE FACTORS 53 56% to 20%, radish rose from 18% to 52% at position B, while the remainder were unchanged. With a further 90° turn in the direction of the arrows (Fig. 1), the radish plant was now opposite the direction of the wind, and nasurtium-Isomeris was toward the wind. Kale at position C to the right side now rose from 20% to 43%, mustard at position D (away from the wind) rose slightly and v/as now 25%, radish at the left side was now 26% and nasturtium-Isomeris (toward the wind) was still relatively unchanged at 2 % . The final 90° turn placed kale in the position toward the wind. Here, 53% of the larvae went to kale, 15% to mustard (right side), 13% to radish (opposite side), and 13% to nasturtium-Isomeris (left side) . CONCLUSIONS: ( 1 ) Of the four sides of the flat which could face the direction of the wind, the plants on that side were favored in three of the four arrangements, namely, mustard, radish and kale. Only when the nasturtium-Isomeris combination v/as toward the wind was the fre- quency in position B not increased. In fact, at that time, it was decreased. (2) These tests, of course, cannot indicate which of the three favored plants are relatively more favored since they each received about the same percentage of selection when toward the wind. However, when away from the wind, radish received fewer selections than either kale or mustard. (3) The Isomeris-nasturtium combination received its highest selection when in position A (left side). In this position also, more larvae left the flat without a plant selection than in any other position. (4) The higher selection of kale (43%) while at position C (right side), than radish (26%) while at position A (left side) seems to indicate a preference for kale rather than for radish. ( 5 ) The lowest selection for the nasturtium-Isomeris combination was when it was in position B (toward the wind). This would seem to indicate that these plants (and Isomeris especially) repel rather than attract larvae of Pieris rapae. (6) The data indicate that the direction of wind movement is of great importance in larval selection of food plant, since in all cases where the plant has a positive effect on selection, that position in the direction of the wind has by far the greatest selective influence. It is apparent that the reverse is also true, that if a plant is repellent, the effect is greatest when the wind comes from that direction. [NOTE: The differences between the figures indicated are so great that the element of chance being involved is negligible. A chi-square test on the first trial gives a probability of less than one in a million that the differences are due to chance alone.] 54 HOVANITZ AND CHANG /. Res. Lepid. EFFECT OF LARVAL SIZE ON PLANT SELECTION Tests for the determination of influence of larval size on plant selection were made in the same manner as indicated previously, except that all tests were carried out in a room with no wind or breezes. Nevertheless, the flat was turned 90° three times for each test. This applies to all subsequent tests for the sake of safety. Two . series of tests were made, using larvae of different origin. The first series of tests (Table 2) was carried out with the use of kale-bred larvae. Larvae of different ages and sizes (size is roughly proportional to age) were separated and measured. Four larvae were isolated, these being the following sizes:: 8.4, 14.7, 19.0 and 21.0 mm in length. The tests were carried out by five trials of a given larva at each position with respect to the direction of the wind (Fig. 1), giving a total of 20 trials for each size larva. TABLE 2. Choice of five kinds of plants by different sized larvae of Pieris rapae ( Series 1 ) Size of Total larvae The number of times and percentage larva goes to Number (mm) Kale Mustard Radish Nasturtium Isomeris None of trials 21.00 16- — 80% 4—20% 0— 0 0—0 0—0 0—0 20 19.00 13—65% 3—15% 3—15% 1—5% 0—0 0—0 20 14.70 11—55% 4—20% 4—20% 1—5% 0—0 0—0 20 8.40 10—50% 5—25% 5—25% 0—0 0—0 0—0 20 One single kale-bred larva was used for each size; tested each larva five times at each position. The 8.4 mm larva selected kale 50%- of the time, radish 25% of the time and mustard 25% of the time. The 14.7 mm larva selected kale 55%, mustard 20%,. radish 20% and nasturtium 5 % of the time. The 19.0 mm larva selected kale 65%, mustard 15%, radish 1 5 % and nasturtium 5 % of the time. The 21.0 mm larva selected kale 80% and mustard 20% of the time and no others. The probability, as caculated by chi-square, that the differences as indicated might be due to chance are less than 1/million, 1/100,000, 1/500, 1/500 respectively for each experiment. CONCLUSIONS: ( 1 ) The increase in selection of kale for each experiment is correlated with increase in size of the larvae ( Fig. 2 ) . (2) Kale is the plant selected over all the others in this experi-\ ment. It should be noted that the larvae had been bred on kale prior to the experiment, and previous generations had been bred on cabbage. (3) Not one larva went to Isomeris'. This is interesting in view of the observation in the preceding experiments that Isomeris was (i);5i-6i, 1962 THREE FACTORS 55 2. The relation between size (and age) of larvae and the percentage of the time kale was selected. The size of the larvae is directly correlated with an increase in selection of kale. repellent to Pieris rapae larvae. (4) The older or larger the larvae, the greater the selection of kale over all other plants ( Tig. 2 ) . The relationship is not a straight- line, but rather a geometric increase. This indicates that there is an increase in the critical powers of food perception in the older or larger larvae. In the Series No. 2 tests (Table 3), larvae were used which had previously been grown on nasturtium. They were obtained from Ventura County on nasturtium and were continued in the laboratory on that plant. Two larvae 22.4 mm in length, and four larvae 16.5 mm in length were used in this experiment. Otherwise, the trials were managed in the same way as in previous experiments with the exception that the total number of trials was 80 for large larvae and I6O for the 56 HOVANITZ AND CHANG /. Res. Lepid. smaller larvae. The selection by these larvae bred on nasturtium (Table 3)' was radically different from the selection by larvae bred on kale ( Table 2 ) . The selections of kale by the small larvae bred on nasturtium were 15% as compared with 55% for a comparable size bred on kale. The selections of kale by the large larvae bred on nasturtium were 22% compared with 80% for a comparable size bred on kale. The greatest shift in food plant selections for the nasturtium-bred small larvae was 49% to mustard as compared with 25%, 9% to nasturtium as compared with 0%, and 16% with no selection as compared with 0%. For the large larvae, comparable data were 34% to mustard as compared with 20%, 33% to nasturtium as compared with 0%, and 9% with no selection as compared with 0%. CONCLUSIONS: ( 1 ) The shifts in food-plant-selection by the larvae grown on different plants are highly significant and indicate a direct relationship between the feeding habits of the larvae, and their selection when given a free choice. (2) The direct relationship between the size (and age) of the larvae and their selection of the plant previously eaten is clearly indicated. (3) A plant previously of nearly negligible selection may be the most highly selected if the larvae have been grown on that plant. For example, larvae selected nasturtium over all others (33%) after being bred on that plant (Table 3), whereas larvae bred previously on kale had almost no selectivity for that plant. INFLUENCE OF PREVIOUS FOOD PLANTS OF LARVAE Additional experiments to test the influence of previous exposure to larval food plants are reported on here. The first comparative test involves two strains of Pieris rapae, one from a population growing in the wild on black mustard and the other growing on cabbage. These strains were kept in the laboratory for several generations; the mustard strain was bred for over six genera- tions on mustard in the laboratory and the kale strain was bred on kale in the laboratory for over ten generations. Larvae were selected for these experiments which had a size of 14 to 16 mm. Using twenty-five larvae of each strain, 600 tests of selection were made (Table 4) : the selected plants were mustard, kale, nasturtium, Isomeris and Cleome. The only significant differences detected in the tests between the two strains were the relative selections made of mustard and kale. Larvae from the mustard-bred strain selected mustard 61% of the time compared with the kale-bred strains of larvae of 24% on that plant. On the other hand, kale was selected by the mustard-bred larvae 20% TABLE 3. Choice of fi¥e kinds of plants by different sized larvae of Pieris rapae ( Series 2 ) 57 1^62 THREE FACTORS . m : > CM d ! ^ Z lb, , m tt ! .2 0 0 jj| * ^ : © » \o ffl cwo 0 1 i z ,u 0 0 01 m © m E 0 II m © ra «0 E E .2 mc\ © u ""C m 1 1 1 ■s c m Z \o ,-4 ffi m © E ^ 0 H- 0 m 1 1 © M E 3 c m *0 C\ i- •S m 1 1 1 1 r- o\ CN r^ fN| \f\ ^ T £ 1 w ^ CM og ^ fr\ m — 6 £ rvivd fNj 1-1 a .S g d r ’S S 3 « o ^ s « 4* y -i o « z S”" o OT aj ^ « 3 9J U Ci > « e ^ a, S « ^ t -G «— « M O *u S 3 ^ a . S g i p o G ■g 3 Si5 d U CN ^ « flj Ms- a rt o > « — I P4 4> •S-a a 2 3® ^ ^9 G ^ m ""W 2 1 11 8-§ « ej 'S ^ bo S e a m O T II u ° g s 8- . uO S-l ^ S p S 2 2 S ^ So «« 0^0 OJ ^ a r-l 2^ S M 3 «J s > ^ fe O ^ ^ ^ g G '-2 b ^ o *-**-• 2 y P«i S 6 8^ The nastuf tram-bred larvae were from Ventura County, collected in a nasturtium garden. The parents of the kale-bred larvae were collected io western Orange County on cabbage; the larva are the first generation fed on kale. 58 HOVANITZ AND CHANG /. Res. Lepid. of. the time as compared with 59% by the kale-bred larvae. Testing these differences by chi-square using the four-fold table indicates a probability of less than 1 /million that the differences could be due to chance alone. This may be calculated in another way, that is, by taking 364 + 354 = 718 for one value and 119 + 144 = 263 for the other. The expectation that these differences might be due to chance alone, that is, that the deviations 718 — 981, and 971 — 263 are not due to 2 2 ‘ chance is well over one in a million. CONCLUSIONS: ( 1 ) The shift in selection of mustard or of kale according to whether or not the larvae were from a kale or mustard strain is indicated by the experiments. The data are highly significant at an incredibly high level. ( 2 ) There is no significant change indicated in the selective level of the other plants involved as to whether or not the larvae were from the mustard or kale strain. The second series of data on the problem of food plant selections involves nasturtium-bred larvae from a nasturtium strain and kale- bred larvae from a cabbage strain (Table 5). The experiments were carried out as before, two larvae of each strain were used, the same size for a total of forty trials each, of which ten were in each position as indicated in Figure 1. The differences between the plant selections are again highly significant. The nasturtium -bred larvae as compared with the kale- bred larvae preferred mustard (35% as compared with 22.5%) and nasturtium (15% over 4%). The kale-bred larvae preferred kale (49% over 12.5%) and radish (22.5% over 14.0%). Neither selected Isomeris to any extent. A high proportion (22%) of the nasturtium-bred larvae rejected all plants. CONCLUSIONS: ( 1 ) As in previous experiments, there was a shift in selection by the larvae according to the previous larval food plants. ( 2 ) This shift in selection was in the direction of the plant they had previously fed on. For example, nasturtium-bred larvae selected nasturtium better than did kale-bred larvae. ( 3 ) Even though nasturtium-bred larvae selected nasturtium better than did kale-bred larvae, mustard was preferred over nasturtium and over kale. (4) The greater selection of mustard over kale bv nasturtium- bred larvae is significant not only in this experiment but also in a previously indicated test (Table 3). In both cases, kale is poorly preferred as compared with mustard when the larvae were nasturtium- bred. (5) Radish is of lower selective influence when the larvae are (i):5i-6i, 1962 THREE FACTORS 59 nasturtium-bred than when they are kale-bred. This was indicated also in previous tests ( Tables 3 and 1 ) . (6) A significantly higher percentage of rejects (22%) are made by the nasturtium-bred larvae than by the kale-bred. This was also indicated in previous tests (Table 3 as compared with 1, 2 or 4). DISCUSSION All three factors, investigated as to their relevance in influencing larval choice of food, have been shown to be important in that regard. These factors are: ( 1 ) Influence of air movemenr. (2) Influence of age (size) of larvae. (3) Influence of food previously eaten by larvae, or strain of larvae. In addition, by virtue of the laboratory set-up used in these experiments, some further conclusions may be drawn which are not completely in accord with some previous conclusions reached by others. One of these further conclusions is with regard to the ability of lepidopterous larvae to "recognize” food plants from a distance. In the present experiments, the larvae were attracted to the plants from a distance of over 120 mm without any difficulty and responded in most cases immediately. Chin (1950) reports that the larvae of the Colorado potato beetle (Leptinotofsa) could not perceive its food plant from a distance farther than 2mm. Gupta and Thorsteinson (I960) tested the olfactory response of the larvae of the diamond backed moth [Plutella macul- ipennis (Curt.)] toward mustard oil odors and found that 5 mm was the maximum distance of response. Dethier (1959) in a study on food plant distribution and density and larval dispersal as factors affecting insect populations, states in relation to field observations that "Larvae find new food-plants by chance alone.” It is presumed that he had reference to the gross search for new food plants, not the closer relationship. In the case of the experiments of Chin, and Gupta and Thorsteinson, the screen test was used. In this, the larvae were placed on a screen above the food plant or other attractive source. In the present experiments, the larvae were below the food plants. This difference could account for a fundamental difference when testing for an odiferous substance heavier than air. Under natural conditions, our observations have always indicated that larvae tend to climb upwards rather than down so that tests involving simulated natural conditions should be best performed with the attraction above rather than below the larvae (Fig. 3). This point will be covered more fully in tests to be described in later papers. In a paper available to us after the tests here indicated were performed, (Takata, 1959), there has been indicated a change in food plant preference of larva when reared successively on different food 60 HOVANITZ AND CHANG /. Res. Lepid. 3. Larva of Pieris rapae on edge of pot during one of the selection experiments reaching out to black mustard plant three to four inches away. The larva had already traveled several inches under directed impulse to get this far. plants for several generations. These results complement ours, even though the technique of testing used by Takata was different. He used cabbage and radish with Pieris rapae crucivora larvae. In testing, he used actual feeding experiments of the weight of leaves of the plants eaten by the larvae. (i):5i-6i, 1962 THREE FACTORS 61 SUMMARY The evidence in this paper indicates ( 1 ) that larvae of Pieris rapae are capable’ of detecting their preferred food plant from a distance of at least 120 mm under the conditions of these experiments, (2) that the direction of wind movement is very important to detection of food plant, presumably through odor, (3) that the older and larger larvae are much better at selecting the preferred food plant at least at a distance and (4) that the influence of food previously eaten by the larvae, or the strain of the larvae, directly influences their choice of food in the direction of the previously eaten food. LITERATURE CITED CHIN, CHUN-TEH. 1950. Studies on the physiological relations between the larvae of Leptinotarsa decemlineata Say and some solanaceous plants. H. Veenman and Sons, Wageningen, Holland, pp. 1-88. DETHIER, V. G. 1959. Food plant distribution and density and larval dispersal as factors affecting insect populations. Canad. Entom. 91: 581-596. GUPTA, P. D. AND A. J. THORSTEINSON. I960. Food plant relation- ships of the diamond-backed moth (Plutella maculipennis (Curt.)). I. Gustation and olfaction in relation to botanical specificity of the larva. Ent. exp. and appl. 3 (I960) :24l-250. TAKATA, N. 1959. Studies on the hobt preference of common cabbage butterfly, Pieris rapae crucivora (Boisduval). VI. Change in the food preference of larvae when reared successively by the definite food plant for several generations (preliminary report). Jap. J. Ecol. 9:224-227. 1(1) :63-71, 1962 Journal of Research on the Lepidoptera 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Copyright 1962 THE GENERIC, SPECIFIC AND LOWER CATEGORY NAMES OF THE NEARCTIC BUTTERFLIES PART 1 — The Genus Pieris PADDY McHENRY 1032 E. Santa Anita, Burbank, California PREFACE INTRODUCTION The author has asked that I write a brief introduction to this, the first of a series of compilations of generic, species and lower category names of the Nearctic butterflies. The objects of the work are (1) to give a complete list of all the names proposed for the butterflies of the Nearctic region from the year 1738 until the date of publication of the list, (2) to indicate the date and place where originally published, (3) to indicate the types of the genera, species and lower category names where known, (4) to indicate the type locality in the original author’s words for types, except those for the generic names, and (3) to indicate the location where types have been deposited where indicated by the author. The biological application of these names so indicated is not to be attempted in this series. Instead, this effort is left to the individual who makes a comprehensive study of the biology of the groups con- cerned, such a person rarely having the time and inclination to make the detailed and laborious studies of nomenclature as such. There is urgent need for a work such as this even in a field as well civilized as the Lepidoptera; in fact, this kind of work is so long overdo that the civilized state is rapidly reverting to chaos. Each new volume which appears from the press on this group of insects plows under well established names with no reason given to justify such destruction. A most valuable start in the direction of making reason out of chaos was started by Francis Hemming in his TThe Generic Names of the Holarctic Butterflies” published in 1934. This work is proceeding also in the sam.e direction. When the various parts of this work are completed, it is expected to reissue the whole as a separate volume. Then, perhaps we can prevail upon Mr. McHenry to go on to a similar effort with the moths, William Hovanitz 63 64 McHENRY /. Res. Leptd. Each entry for the generic names includes the following; the name, the author’s name, the most accurate determinable publication date, the bibliographical reference (including the volume or subvolume for the date indicated ) , the type species, the selection of the type species and the supporting data to justify the type indicated. Each entry for the specific names includes the following; the name as originally given, the author’s name, the most accurate determinable publication date, the bibliographical reference, data on type locality, sex, season, etc., other spellings of the name and notations of any homonym. The date of publication of a name is indicated in a manner that shows its relative accuracy; for example, "mailing” and "release” dates if known are given preference over "publication” dates if they are not the same. All dates are given in English and normally in the sequence of day, month and year. Where more than one date is indicated, the more accurate one is given after the less accurate one. Dates ascertained directly from a source whose purpose was the determination of time of publication of a work are given without brackets; others, ascertained less directly are given in brackets and explained in a footnote. Combinations of dates with and without brackets may be given when necessary. I have seen originals or photostats of all references except as noted but have been unable to ascertain the contents and dates for the various parts of Verity’s Rhopalocera Palaearctica. The names of the genus Pieris are listed under the five species indicated, since in this genus the separation does not involve great difficulty. I have been aided in this regard by William Hovanitz, or by reference to the original description. Should these be in error, no harm is done since the total list, not the biological interpretation of the categories to which the names applied, is the purpose of the listing. Should anyone find names which have been inadvertently omitted from this listing, a great service will be rendered to so notify the author or the editor of this fact so that corrections may be made. I take this opportunity to express my gratitude for the use of facilities and to the staffs at the Allan Hancock Foundation Library of the University of Southern California, and at the Los Angeles County Museum Library & Entomological Department. A debt of gratitude is tenderd to the photographic service sections of the U. S. Department of Agriculture Library, Harvard College Library, New York Public Library, University of California Library and Cornell University Library. An extended exchange of correspondence with Mr. C. F. DosPassos has been most helpful. LIST OF GENERIC NAMES USED OR AVAILABLE FOR PIERIS ASCIA Scopoli. Type; monuste L. (i):63-7I, 19^2 NAMES Pt. 1 65 MANCIPIUM Hubner. Type: hellica Hubner = hellica L. = helice L. PIERIS Schrank. Type: hrassicae of var. authors = brassicae L. ANDROPODUM Hubner. Type: brassicae Hubner = brassicae L. GANORIS Dalman. Type: brassicae L. TACHYPTERA Berge Type: brassicae L. PONTIA Fabricius. Type: daplidice Fab. = daplidice L. SYNCHLOE Hubner. Type: callidice Butler = callidice Esper = callidice Hubner ASCIA SCOPOLI. 1777. Introd. Hist. Natur. : 434, no. 175. Includes among others no. 80 (="P. D. [C.] monuste” of Linnaeus, 1767) which in turn is Papilio D. C. monuste Linnaeus, 1764. Type Papilio D\anaus\ C[andidus]. monuste Linnaeus. 1764. Museum Ludovicae Ulricae ([I]) : 237, no. 56. Type Selection. Scudder. 1872 [in or pre-June] L Syst. Rev. Some Amer. Butt., p. 40. Says of Ascia: "Type Papilio monuste Linn.” MANCIPIUM HUBNER. [1 Jan. 1807] - [19 Dec. 1807]L Samrn. Exot. Schmett, 1: pi. [141], ff. 1-4. Gives "Mancipium vorax hellica”. As used by Hubner, hellica is considered to be hellica Linnaeus, 1767 which in turn is helice Linnaeus, 1764. Type. Papilio D[anaus]. C[andidus]. helice Linnaeus. 1764. Museum Ludovicae Ulricae ([I]) : p. 243, no. 62. Type Selection. Hubner. As above. Gives only hellica, at the time, which becomes the type. PIERIS SCHRANK. 1801. Fauna Boica 2(1): 152, no. 198; 2(1): 160-170, nos. 1283-1296. Includes Pieris brassicae which he says is Papilio brassicae of Goeze, Borkhausen, Brahm, and Esper. Of these authors, brassicae is considered to be hrassicae Linnaeus, 1758. Type. P\ap)ilio\ D\anaus\ [Candidas] hrasicae Linnaeus. 1758. Syst. Nat. 10th Ed. 467-468, no. 58. Type Selection. Latreille. 1810. Consid. Gen. Anim. Crust. Arach. Ins. p. 440. Says: "Pieride. Pontia brassicae, Pab.”. As used by Latreille, brassicae Fabricius !s considered to be brassicae Linnaeus, 1758. ANDROPODUM HUBNER. 1822. [post 22 Sept.]." Syst.-Alph. Verz. pp. 2-5, 7-9. Includes "Brassicae L, 401-403. Andropodum vorax.” Type. P\apilio\ D[anaus]. [Candidas] hrassicae Linnaeus. 1758. Syst. Nat. 10th. Ed. 1:467-468, no. 58. Type Selection. Hemming. Sept. 1933. Entomologist: 66(844) : 199. Says of Andropodum: "Type =: Andropodum hrassicae Linn.” GANORIS DALMAN. 1816. Kongl. Vetenskaps Academiens Handlingar 1816(1) : 61-62, no. 8; 1816(1) : 86-90, nos. 1-10; tablet 1, no. 8; tablet 2, no. 8. Includes G. brassicae in which he cites Linnaeus Fauna Sv., pp. [269] -[2701, no. [10351 (this is hrassicae Linnaeus, 1761, which is hrassicae Linnaeus, 1758). Type. Plapilio]. D[anaus]. [Candidas] Brassicae Linnaeus. 1758. Syst. Nat. 10th Ed. 1:467-468, no. 58. Type Selection. Dalman. As above. Says of Ganoris: "Generis Typus: Pap- hrassicae”. 66 McHENRY /. Res. Lepid. TACHYPTERA BERGE. 1842. Schmetteclingsbuch, p. 19, pp. 92-105. Not seen, given as per Hemming, 1934. Type. P[apilid\, .Dlanaus\ [Candidus] brassicae Linnaeus. 1758. Syst. Nat. 10th. Ed. 1:467-468, no. 58. Type Selection. Hemming. Feb. 1934. Entomologist: 67(849) :38. Says of Tachyptera: "Type Papilio brassicae Linn., 1758.” PONTIA FABRICIUS. 1807. In Illiger. Magazin fur Insektenkunde. 6: 283, no. 23. Includes: ’Tap . , . Daplidice”. As used by Fabricius daplidice is considered to be daplidice Linnaeus, 1758. Type. P[apilio]. D[anaus]. [Candidus] daplidice Linnaeus. 1758. Syst. Nat. 10th Ed. 1:468, No. 2. Type Selection. Curtis. 1824. Brit. Entom. 1: pi. 48. Not seen, given as per Hemming, 1934. SYNCHLOE HUBNER. 1818 [post 22 Dec.]^. Zutr. Samm. Exot. Schmett. 1st. 100 (text in 1): p. 26, under no. 76. (Synchloe autodice). Includes "S. callidice”. Refers to his earlier publication of this butterfly. Type. Papilio callidice Hifbner. [24 Dec. 1799] - [13 Apr. 1800]^. Samm. Europ. Schmett. (Papiliones) : plate 81, figs. 408-409. Page 63, no. 5 and plate 108, figs. 551-552 were published later. Type Selection. Butler. 12 Sept. 1870. Cistula Ent. 1(3):51. Says of Synchloe: "Type S. callidice Esper”. As used by Butler callidice Esper is considered to be callidice Hubner [24 Dec. 1799] - [13 Apr. 1800]. LIST OF SPECIES AND LOWER CATEGORY NAMES USED OR AVAILABLE FOR PIERIS 1. P. (PONTIA) BECKERII EDWARDS. heckerii Edwards. gzmderi Ingham. pseudochloridice McDimnough. 2. P. (PONTIA) SISYMBRII BOISDUVAL. elivata ( B. & B. ) . jlava Edwards. flavitincta Comstock, J. A. sisymhrii Boisduval. transversa {B. 8iP>.) 3. P. (PONTIA) PROTODICE BOISDUVAL & LECONTE calyce Edwards. . nastuftii Edwards ( Bdv. Ms. ) . nehoni Edwards. occidentalis Reakirt. protodice Boisduval & LeConte. vernalis Edwards. 4. P. (PIERIS) NAPI (LINNAEUS). acadica Edwards. aestiva Edwards. arctica Verity. borealis Grote. casta Kirby. i(i):6}-7T, 1962 NAMES — Pt. 1 67 castoria Reakirt. cottlei Gunder. cruciferamm Boisduval. flava Edwards. flava Edwards. ffigida Scudder. hulda Edwards. hyemalis Edwards. iberides Boisduval. macdunnoughii Remington. marginalis Scudder. micro striata Comstock, J. A. mogollon Burdick. napi (Linnaeus). nasturtii Boisduval. ochsnheimeri Staudinger. oleracea Harris. pallida Scudder. pallidissima B. & McD. pseudobryoniae Verity. pseudoleracea Verity. pseudonapi B. & McD. resedae Boisduval. venosa Scudder. virginiensis Edwards. 1870. virginiensis Edwards. 1881. 5. P. (PIERIS) RAPAE (LINNAEUS). aestivus Verity. immaculata De Selys-Longchamps. immaculata Cockerell. immaculata Skinner & Aaron. metra Stephens. novangliae Scudder. rapae ( Linnaeus ) . yreka Reakirt. 1. PIERIS (PONTIA) BECKERII EDWARDS. heckerii Pieris Edwards. [Sept. 1871]*^. Butt. N. Amer. 1(8) : [28a] - [29]; plate [81], figs. 4-7. 'cf described. "Virginia City, Nevada, April 1870.” "Four individuals”. McDunnough, 1 Dec. 1928, spells heckeri. gunderi, Ascia heckerii Ingham. 14 Apr. 1933. Pan-Pacific Ent. 9(2) :75. Holotype ? : "Bouquet Canyon, Los Angeles County, Calif”. "Bred from larva”, "emergence date June 21, 1932”. Type deposited: Calif. Acad. Sci., San Francisco, Calif. pseudochloridice, Pieris heckeri McDunnough. 1 Dec. 1928. Can. Ent. 60(11) : 266-267. Holotype d* : "Oliver, B[ritish]. C[olumbia]., April 24 . . . No. 2861 in the Can. Nat. Coll., Ottawa”. Allotype ? : "Oliver, 68 McHENRY /. Res. Lepid. B.C., April 22”. Paratypes — I d*: "Hedley, B. C., May 15”. 1 9: "Osoyoos, B.C.” 1 + : "Osoyoos, B.C.” 2. PIERIS (PONTIA) SISYMBRII BOISDUVAL. elivata, Ascia sisymhni Barnes & Benjamin. 8 Dec. 1926. Bull. Sou. Calif. Acad. Sci. 25(3) :88, no. 33a. "Type locality: Glenwood Springs, Colo.”. "Number and sexes of types: Holotype d* May 1895, Allotype ? May 1895, 7 d» 14 9 Paratypes, various dates April to June”. jlava, Pieris sisymhri Edwards. Apr. 1883. Butt. N. Amer. 2(11): [67]; plate [151. fig 5. ? described. No locality, series data nor dates given. flavitincta, Pieris sisymhrii Comstock, J. A. 20 Feb. 1924. Bull. Sou. Calif. Acad. Sci. 23(1) : 19-20; plate 7, fig. 9. 9 described. "The example before us was captured . . . April 30th, 1911 at Cranbrook, British Columbia”. Intends to propose new name for jlava Edwards. sisymhrii, Pieris Boisduval. 1852 [25 Feb. - 22 Dec. Ann. Soc. Ent. France. 2nd Ser. 10 (2) :284, no. 8. 9 described. [California]. No series data nor dates given. Edwards, Apr. 1883, spells sisymhri. transversa, Ascia sisymhrii Barnes & Benjamin. 8 Dec. 1926. Bull. Sou. Calif. Acad. Sci. 25(3): 88, under no. 33. "Type localities: Paradise, Cochise Co., Ariz.; Redington, Ariz.”. "Number and sexes of types: Holotype cf , Allotype 9,1 9 Paratype, all March; 1 cf Paratype, no date”. 3. PIERIS (PONTIA) PROTODICE BOISDUVAL & LECONTE. calyce, Pieris Edwards. Nov. 1870. Trans. Amer. Ent. Soc. 3(?) sign. 25: 189, no. 1 (Nomen nudum); 189. d* described. "From Nevada; in the collection of Henry Edwards”. No series data nor dates given. nasturtii, Pieris Edwards (Boisduval Ms. name). [9 May 1864]°. Proc. Ent. Soc. Phila. 2(4) : 501, no. 1 (Nomen nudum); 501. d* & 9 described. "San Francisco; from Dr. Behr, who informs me that it is common in some localities in the vicinity of that city”. No series data nor dates given. nelsoni, Pieris Edwards. Apr. 1883 Butt. N. Amer. 2(11): [71] - [72]; plate [15], figs. 6-7. "1 cf ... St. Michael’s Alaska, June, 1881”. occidentalis, Pieris Reakirt. June 1866. Proc. Ent. Soc. Phila. 6 (pp. 121-152) : 133-134. cf & 9 described. "Hab. - Rocky Mountains, Colorado Territory, California, (Coll. Tryon Reakirt)”. No series data nor dates given. protodice, P[ieris]. Boisduval & LeConte. 1829. Hist. Gen. Icon. Lepid- Chen. I’Amer. Sept. 1 (5) :45-46. Plate 17, figs. 1-3 published later. A & 9 described. "Elle parait au printemps et a la fin de juin aux environs de New-York. Nous en avons vu un individu pris dans le Connecticut”. vernalis. Pieris Edwards. [9 May 1864]®. Proc. Ent. Soc. Phila. 2(4) : 501, no. 2 (Nomen nudum); 2 (4) :501-502. cf & 9 described. "In the collection of Mr. Geo. Newman and Mr. Wilt are several specimens, taken, as I am informed, at Red Bank, New Jersey, in the month of May”. "There is also in the Society’s collection a pair from the Rocky Mountains, that appear to be identical with it”. 4. PIERIS (PIERIS) NAPI (LINNAEUS). _ . acadica, Pieris napi Edwards. June 1881. Papilio l(6):86-88 (in pt.), sub no. 1; p. 98, no. 1 of no. 3; p. 99; plate 3, figs. 10-11. It appears that one pair was from Southern Newfoundland and was taken in the last week in July, it also appears another set ( 1 cf 2 9 ) had emerged from pupae in August and was from the same locality. aestiva, Pieris napi oleracea Edwards. June 1881. Papilio l(6):86-88 (in pt.. Nomen nudum), sub no. 1; l(6):89-95 (in pt.) sub no. 3; l(6):95-98 (in pt.), sub no. 4; 1(6):98, no. 3 of no. 3; i( 1962 NAMES — Pt. 1 69 1(6):99; plate 3, figs. 15-16. cf & described. "NewYork” (see figs, explanation, p, 99). "summer brood”. No series data given. arcUca, Pieris napi frigida Verity. 1905-1911. Rhopal. Palaearctica : pp. 333-334 (in pt., cites figs 32-33 on plate XXXII and figs. 16-17 on plate LXVII as P. n. f. arctica)\ plate XXXII, figs. 32-33; plate XXXII explanation page, figs. 32-33 (gives as P. n. frigida); plate LXVII, figs. 16-17; plate LXVII explanation page, figs. 16-17 (gives as P. n. arctica); p. xxviii (cites figs. 31-33 & 36-37 on plate XXXII and figs. 16-17 on plate LXVII, gives as P. n. /. arctica). "31 . . . cf (ile d’Yesso, Japon)”. "32 . . . c? , (Norvege sept.)”. "33 . . . d* (Nulato, Alaska)”. "36 ... $ (Nulato, Alaska)”. "37 ... 9 .(Finmark, Scandinavie) ”. ”16 , 9 (Saltdalen, Norvege sept.) ”. "17 . . . ? (Laponie)”. Figs. 31, 33, 36: [coll, de Joannis]. Fig. 32: [coll. Stefanelli]. Fig. 37: [coll. Obth.] Fig. 16: le coll. Mourray]. Fig. 17: [e coll. Leach]. No series data nor dates given. borealis, Ganoris oleracea Grote. Nov. 1873. Bull. Buffalo Soc. Nat. Sci. 1(4) sign. 24:185. "months of June and July, on the Island of Anti- costi”. No series data nor sex given. casta, Pontia Kirby, W. 1837. In J. Richardson. Fauna Boreali-America (4) :288, no. 1; plate 3, fig. 1. "Three specimens taken in Lat. 65°”. No sex given. castoria, Pieris Reakirt. [11 June 1866] - [13 Feb. 1867]*^. Proc. Acad. Nat. Sci. Phila. [18] (3) :238, no. 2. d* described. "Hab. - California. Coll. Tryon Reakirt”. No series data nor dates given. cottlei, Pieris napi castoria Gunder. 6 July 1925. Ent. ^ews 36(7) :197- 198, no. 9; plate 5, fig. 9; Holotype d : "(author’s Coll.), Anderson Springs, Lake County, California, May 5, 1919”. crucifer arum, Pieris Boisduval. 1836 [in or post Apr.]^°. Hist. Nat. Ins. Spec. Gen Lepid. 1 (text in 1) :519, no. 119. d* described, "dans les provinces septentrionale des Etats-Unis”. No series data nor dates given. flava, Pieris napi pallida Edwards. June 1881. Papilio 1(6) :89”95 (in pt., unnamed), sub no. 3; 98, sub no, 2 of no. 3. "Washington Terr[itory].”. "One of these [a $ ] is yellow on the upper side”. No date given. flava, Pieris napi venosa Edwards. June 1881. Papilio 1(6) :88-89 (in pt,, unnamed), sub no. 2; 98, sub no. 1 of no. 2. Fig. 7 of plate 2 (although not so named in plate explanation, p. 99) appears to be flava. "A large percentage of female Venosa are yellow on the upper side”. No locality, series data nor dates given. frigida, Pieris Scudder. Sept. 1861. Proc. Boston Soc. Nat. Hist. (8(?) sign. 12:181-182. "Two . . . males . . . two females”, "on Caribou Island, Straits of Belle Isle”. No dates given. hulda, Pieris Edwards. Sept. 1869. Trans. Amer. Ent. Soc. 2(?) sign. 48:370. "From Kodiak, 1 o* . Coll, of Henry Edwards”, No date given. hyemalis, Pieris napi oleracea Edwards. June 1881. Papilio l(6):88-89 (in pt.. Nomen nudum), sub no. 2; 89-95 (in pt.), sub no. 3; 95-98 (in pt.), sub no. 4; 99; plate 2, fig. 8. d* &■ $ described. Directly and indirealy cites material from New Hampshire, Massachusetts, Lake Superior area and Mt. Hood, Oregon. No series data nor dates given. iherides, Pieris Boisduval. [1869, pre 1 Nov.]^’. Ann. Soc. Ent. Belgique 12 (in 1):39, no. 9. d* & ? described. [California]. No series data nor dates given. macdunnoughii, Pieris napi Remington. 17 Sept. 1954. The Lepid News 8(3-4) :75. A new name for Pieris napi pseudonapi Barnes & McDun- nough, 5 Dec. 1916. See pseudonapi for series data. marginalis, Pieris Scudder. Stpt, I86I. Proc. Boston Soc. Nat. Hist. 8(?) sign. 12:183. "(Id* ,1 9 ) which are in the Museum of Comparative Zoology”. The male . . . from Gulf of Georgia, and the female from Crescent City, California”. No dates given. 70 McHenry /. Re$. Lepid. mkrostfiata, Pieris napi Comstock, J. A. 12 Sept. 1924. Bull. Sou. Calif. Acad. Sci. 23(4) :125. Holotype f:- "Eldredge, Sonoma County, Cali- fornia. March 13, 1911”. Allotype [•$■]: “same locality and on the same date”. 'Taratype; one male - same locality and date”. mogollon, Pieris napi Burdick. 28 Sept. 1942. Can, Ent. 74 (8) :154-155. Holotype : "Mogollon Range, Catron Co., New Mexico, 5»7-40. No. 5222 in the Can . . . Nat . , . Coll . . . , Ottawa”. Allotype $ : "same data”. "Paratypes. 2 ^ ,same data; 3 cf , Sierra Blanca, Lincoln Co., N. M., 5-25-40; all in the Can , . . Nat . . . Coll . . . , Ottawa. 1 d* , Mogollon Range, Catron Co., N. M., 5-7-40 and 1 $ , Sierra Blanca Range, Lincoln Co., N. M., 5-20-40; in the U , . . S . . .^at . . . Mus . . Fifty-six paratypes from the above locations in the collection of the author, from which other museums will be supplied”. napi, P[apilio]. Dlanausl. [Candidus] Linnaeus. 1758. Syst. Nat. 10th. Ed. 1 ;468, no, 60. No locality, series data, sex nor dates given. Edwards, June 1881, spells napae. nasturtii, Pieris Boisduval. [1869, pre 1 Nov.]^L Ann. Soc. Ent. Belgique (12) (in 1) :38-39, no. 7. described, "dans les champs decouverts au pied de la Juba”. "Nous o’avons vu que des males”. No series data dates given. A homonym of Pieris nasturtii Edwards, [9 May 1864]. ochsenheimeri, Pieris Staudioger. End Apr. 1886. Ent. Zeitung Stettin 47 (4-6) : 199-200. d & 9 described, "wo sie die beiden Haberhauer in Anzahl Ende Juni bei Namagao, jedenfalls hoch in den Gebirgen ge fangen batten. Auch Maurer sandte mir in demselben Jahre ein im Alai-Gebirge (sudlich von Margelan) gefangeoes + ein”. No series data given. oleracea, Pontia Har^fs. 10 July 1829. The New England Farmer & Horti- cultural Jour. 7(51) :402. "Habitat ... in New Hampshire, and Massa- chusetts”. No series data nor sex given, "two broods in a season”: [May- June - August]. pallida, Pieris Scudder. Sept. 1861. Proc. Boston Soc. Nat. Hist. 8 (?) sign. 12:183. "five specimens (3 d* » 2 9.), which are in the Museum of Comparative Zoology”. "Gulf of Georgia”. No dates given. pallidissima, Pieris napi Barnes & McDunnough. 5 Dec. 1916. Contrib. Nat. Hist. Lepid N. Amer. 3(2) :59; 142; plate 6, figs. 4-5 & 10. "second generation (July, August)”, "type cf and $ from Provo, Utah”. No series data given. pseudobryoniae, Pieris napi frigida Verity. 1905-1911. Rhopal.- Palaearctica : p. 146 (cites figs. 46-37 on plate XXXII, gives as P. n. f. pseudo- bryoniae); plate XXXII, figs. 36-37; plate XXXII explanation page, figs. 36-37 (gives as P. n. frigida); p. xxviii (cites only fig. 37, gives as P. n. f. arctica pseudobryoniae) . "36 ... 9 (Nulato, Alaska) [coll, de Joannis]”. "37 . . (Finmark, Scandiriavie) [coll. Obth.]”. No series data nor dates given. pseudoleracea, Pieris napi frigida Verity. 1905-1911. Rhopal. Palaearaica: p. 146; plate XXII, fig. 38; plate XXXII explanation page, fig. 38; p. xxviii. "38 . . . d' (Labrador) [e- coll. Boisduval in coll, obth.]”. No series data nor dates given. pseudonapi, Pieris napi Barnes & McDunnough. 5 Dec. 1916. Contrib. Nat. Hist. Lepid. N. Amer. 3(2) :57; 142; plate 6, figs. 1-3. "type ^ and 9 taken at Silverton, Colo ... in the last week of July”. No series data given. A. homonym of Pieris melete pseudonapi Verity, 1905-1911. resedae, Pieris Boisduval. [1869, pre 1 Nov.]^ ’ Ann. Soc. Ent. Belgique 12 (in 1) :38-39, no. 8. 9 described. "Nous ne connaissons cette espece que par un seul individu 'femelle pris . . . sur les bords du Sacramento”. No date given. venosa, Pieris Scudder. Sept. 1861. Proc. Boston Soc. Nat. Hist. 8(?) i(i):6}-ji, 1^62 NAMES — Pt. 1 71 sign. 12:182-183. "twenty specimens (5 cf , 15 9 ), brought to the Museum of Comparative Zoology, from San Mateo and Mendocino city, California”. No dates given. vkginiensis, Pieris Edwards. Jan. 1870. Trans. Amer. Ent. Soc. 3(?) sign. 2:10, no. 5 (Nomen nudum); 13-14. cf & $ described. "Not uncom- mon in the Kanawha district [West Virginia] in the month of May, and there replacing Oleracea. I have received from Mr. Saunders occasionally specimens taken by him at London, Canada”. No series data given. virginiensis, Pieris napi oleracea hyemalis Edwards. June 1881. Papilio 1 (6) :95-98 (in pt.), sub no. 4; 98, under no. 2 of no. 2. " . . . Virgin- iensis . . . has become a true species, although unquestionably, in a higher latitude, it appears as an occasional aberration only of Oleracea". Speci- men data is obscure. A homonym of virginiensis Edwards, Jan. 1870. 5. PIERIS (PIERIS) RAPAE (LINNAEUS). aestivus, Pieris rapae Verity. 16 May 1913. Jour. Linn. Soc. (London), Zool. 32 (215) :177-178. "summer brood”. No locality nor series data given. immaculatag Pieris rapae De Selys-Longchamps. 1857. Ann. Soc. Ent. Belgique 1 (in 1) :5, under no. 5. [Belgique]. No series data, sex nor dates given. immaculata, Pieris rapae Cockerell. Apr. 1889. Entomologist: 22(311) :99. No locality, series data, sex nor dates given. Cites p. I6l in Newman’s Brit. Butt. A homonym of Pieris napi immaculata De Selys-Longchamps, 1857. immaculata, Pieris rapae Skinner & Aaron. 2 July 1889. Can. Ent. 21(7) : 128-129. "five specimens in the collection of Am. Ent. Soc., Dr. Skin- ner and E. M. Aaron”. No sex nor dates given. A homonym of Pieris rapae immaculata De Selys-Longchamps, 1857 and Cockerell, Apr. 1889. metra, Pontia Stephens. 1 Aug. 1827. Illust. Brit. Ent. 1(?) sign, d: 19-20, no. 4:146-147 published later, cf & ? described. "The insect occurs early in April and a second time towards the end of June. I obtained specimens of the first brood at Hertford; and of the second I captured some this season, at Ripley at the latter period”. novangliae, Ganoris rapae Scudder. Apr. 1872. Can. Ent. 4(4) :79. "Two of my specimens were taken in June and July, 1869 — - one ... in Shelbourne, N[ew]. H[ampshire] . . . the other ... in Waterville, M[aine]., and all are males”. rapae, P[apilio]. D\anaus\ [Candidusi Linnaeus. 1758. Syst. Nat. 10th. Ed. l(in 1):468, no. 59. No locality, series data, sex nor dates given, yreka, Pieris Reakirt. [11 June 1866] - [13 Feb. 1867]'^. Proc. Acad. Nat. Sci. Phila. [18](3):238, no. 1. cf & 9 described. "Hab. - California. Coll. Tryon Reakirt”. No series data nor dates given. lAmer. Nat. 6: 3S4-559. Work is reviewed in June 1872 number. ^Hemming, 1937. Hubner 1:327-437. ^The work title page signature date qualified by preface signature date on page vi. •^Title page signature date qualified by preface signature on page 6. ■^Hemming. 1937. Hubner. 1:146-324. ^Hemming. 1931. Proc. Ent. Soc. London. Ser. A. 6: 42-45. Gives actual dates of publication for the parts of Edwards’s Butt. N. Amer., (Ser. I). 7Tome 10 title page signature date qualified by the meeting date of 25 Feb. (p. 275) and the evidence indicating that the tome to page 489 was published before the meeting of 22 Dec. (p. Ixxxii in Bulletin). SProc. Ent. Soc. Phila. 1864. 3: p. 695. Notes that No. 4 of Vol. 2 of Proc. Ent. Soc. Phila. was received at 9 May 1864 meeting. 9lndex Scient. Cont. Jour. & Proc. Acad. Nat. Sci. Phila., 1812-1912: pp. xii-xiii. Gives dates of receipt for various parts of the early volume of the Proceedings. 1 OTome I title page signature date qualified by preface date (p. xii). ' 1 Trans. Ent. Soc. London 1869; p. xix (in Proc.). Notes receipt of Tomes 1-12 of the Ann. Soc. Ent. Belg. at 1 Nov. meeting. Tome 12 title page signature date: 1868-1869. I assume the date is fixed on the publication of Tome 12, I have no data for the separate except that its title page signature date is 1869. Journal of Research on the Lepidoptera 1(1) :73-83, 1962 1140 W. Orange Grove Ave., Arcadia, California, U,S.A. Copyright 1962 THE DISTRIBUTION OF THE SPECIES OF THE GENUS PIERIS IN NORTH AMERICA’ WILLIAM HOVANITZ = ’ 1140 ]F. Orange Grove Ave., Arcadia, Calif. The North American species of the Genus Pieris are distributed from the Arctic Ocean to Guatemala. However, none of the species of the genus (sensu stricto) exist in habitats which would be con- sidered as tundra or arctic-alpine on the one hand, or as wholly tropical on the other. To this extent, the genus compares somewhat with Colias; the latter, however, contains two species that do exist in the tundra habitat. The concept of a species which has been applied here is the same as that which has been applied previously to the species of the genus Colias in drawing the distribution maps for that genus (Hovanitz, 1950). This concept is based upon the consideration that the most important factor in stability of a species is the genetical population — or the interbreeding unit. Since, however, genes and their combinations cannot be seen, nor analyzed except by means of their effects on the appearance and physiology of the individuals carrying them, they must be studied through a comparative study of their morphological, anatom- ical or physiological characteristics. Thus, a name applied to one population may be applied to another population, or to many popula- tions, if their characteristics indicate that a similarity exists between the hereditary makeup of such different populations. Such a name may cover quite a diversity of differences in the same population, or even a diversity of different populations if they are genetically separated one from the other by geographical or ecological means. The extent of gene interchange as well as the evaluation of the isolation mechanism determines whether or not one population should be considered as the ’Aided by a grant from tbe National Science Foundation, Washington, D.C. 2The author wishes to thank the following for having made this distributional work possible by making available the collections under their care: C. B. MacNeil, California Academy of Sciences, San Francisco; L. M. Martin, Los Angeles County Museum, Los Angeles; T. N. Freeman, Entomological Research Institute, Ottawa; F. Rindge, American Mus'^um of Natural History, New York; Harry Clench, Carnegie Museum, Pittsburgh; W. S. Field and J. F. Gates Clark, U.S. National Museum, Washington, D.C.; and the Dept, of Entomology, Chicago Natural History Museum, Chicago, Illinois. 3The base maps used in this paper are from Goodes Series of Base Maps by the University of Chicago Press, used by permission. 73 74 HOVANITZ /. Res. tepid. ’same species” as another, or a "subspecies” or "geographical race” of. it, or should not be considered significantly different. These criteria will be more exhaustively dealt with in a later paper. For practicability, different criteria must be used at different times and in varying circum- stances. The ultimate taxonomic designation is a product of weighing a great number of problems, some factors being known and others unknown, with the hope that the best result may be achieved. Another general rule underlies these studies; namely, that stability and uni- formity are desirable in nomenclature, and that past taxonomic decisions should not be upset by hasty, poorly-thought-out changes. But the past should not remain immutable when the decision to change is demanded by sound study or experimental evidence. Implicit in all taxonomic studies is perhaps an amount of individual opinion, perhaps greater than in many more experimental sciences, but even this can be greatly reduced by conscious effort to separate fact from opinion. It is the belief of the author that all facts upon which decisions are based should be stated; these are the "data” which can be refuted. New decisions should be made then only upon presentations of new data. Taxonomy by "hunches” is to be avoided; decisions should be accompanied by evidences for and against any decisions made. THE SPECIES OF PIERIS There are five clearly defined species of Pieris in North America. Of these, only four were native prior to 1800. Besides these, there are two other groups of related populations which might be considered "species,” or sibling species, and a large number of geographical varia- tions within each of these species, many of which are distinctive enough to be called "geographical races” or subspecies. Division of the species into geographical races is not to be considered in this paper, with the exception of the two sibling species just mentioned. Pieris rapae. This is one of two common European species of Pieris, the first immigrants* of which appear to have been introduced into North America prior to I860. Since that time, the species has spread over the whole of the Unitd States and much of Canada and Mexico, wherever the natural vegetation has been at least partially replaced by European weeds, such as black mustard, or by cruciferous agriculture. This species has a wide range of habitat which is suitable for its existence, since it is found from the warm parts of the Gulf of Mexico nearly to the tundra of Canada, and from the Imperial Valley below sea level to high mountain meadows in the Sierra Nevada or the Rocky Mountains. The food plants utilized are a variety of cruciferous plants, among which are all varieties of cruciferous garden vegetables (cabbage, kohlrabi, kale, etc.) as well as many weedy plants such as mustards, radish, sisymbrium, etc. Some garden flowers such as nasturtium are also utilized. In view of the "weedy” nature of this species, no map is given to illustrate its distribution. i(i):73-S3, 1962 PIERIS DISTRIBUTION 75 Pieris napi (Fig. 1). This species is found throughout northern Europe and Asia where there is in effect a continuity with the forms of North America across the Bering straits and the islands of the Bering Sea, This is the most northern of all of the species of Pieris, its range extending to the edge of the tundra or slightly beyond along its northern limits. The northern distributional limit from the Macken- zie River and east follows closely the northern limits of the tree line. The species exists along both the Pacific and Atlantic coast lines, but extends farther southward on the Pacific Coast than on the Atlantic Coast. The southern-most population on the immediate Pacific Coast line is at Lopez Canyon, San Luis Obispo County, California at about 35° north latitude. On the east coast at sea level, the most southern locality known is along the coast of New York and Connecticut opposite Long Island at 42° north latitude. However, at increasing elevations in the Appalachian Mountains southward, P. napi ( or P. vifginiensis) may be found as far south as the Great Smoky Mountains of Tennessee and North Carolina at about 35° north latitude. The southernmost extension of range of Pieris napi in the strict sense is in the vicinity of Connecticut and the Appalachians of Vermont, New Hampshire and Massachusetts. Westward, the range passes through the lower part of Ontario, Michigan, Wiconsin, Minne- sota, Manitoba, Saskatchewan and Alberta, always staying north of the prairies. A fuller discussion of the relation of Pieris napi to P. virginiensis is to be had in a separate paper to follow this one. Where the range of P. napi reaches the Rocky Mountains in central Alberta, the species extends clear to the Pacific coast and southward in the Rocky Mountains at increasing elevations to southern New Mexico and Arizona. The species appears to be absent in the drier ranges of the Great Basin and northward in the sagebrush country into south-central British Columbia. Elsewhere, it is widely distributed in the habitats that it prefers, these being damp wooded areas with partial shade, and temperatures not over seventy degrees Fahrenheit.. Of interest in the search for extensions of range are the following specific locations on the southern parts of the range of Pieris napi: California: Coast — - San Luis Obispo County, Lopez Canyon, December through May. California: Western side Sierra Nevada, Merced County, Yosemite Valley, 3-4000 feet, April-May. California: Eastern side Sierra Nevada, Placer County, 6-7000 feet, June. Nevada: No records at present. Arizona: Apache County, White Mountains, August; Gila County, "Globe” July. New Mexico: Otero County, James Canyon, Cloudcroft, August, Catron 6, Mogollon, May, August. Georgia: No records. 76 HOVANITZ /. Res, tepid. FIGURE 1. Map showing the North ndpi. American distribution of Pieris (Pieris) (i):73-^3, 1962 PIERIS DISTRIBUTION 77 South Carolina: No records. North Carolina: Black Mountains, April. Tennessee: Great Smoky National Park, Cherokee, April. Some other points of interest in the distribution of this species are the following records : Pribilof Islands, Alaska, July. Mouth of the Mackenzie River, North West Territories. Great Bear , Lake, N.W.T. Churchill, Manitoba ( Hudson Bay ) . James Bay, Hudson Bay, Quebec. Ramak, Labrador Pieris napi is the only known butterfly to inhabit the Aleutian Island^ or the islands of the Bering Sea. This ability is probably correlated with its adaptation to existence at temperatures below 70 °F, with low solar radiation and high humidity. Pieris (Pontia) protodice (Fig. 2). This is the most widely distributed species of Pieris in North America. Some prefer to place this species, together with the following two species, in a separate genus Pontia for which there is much to recommend. To do so, however, would reduce the size of the. genus Pieris beyond what would seem reasonable, and, therefore, to obviate this difficulty, and yet show phylogenetic relationship, Pontia would best be retained as a subgenus name. Pieris protodice is best adapted to habitats with more sunshine than is required by P. napi. It also is enabled to survive in habitats much too warm for that species. However, cold does not seem to be a limiting factor as long as there is a short warm season and plenty of sunlight. These conditions, to some degree, pertain to the area occupied by the species. It apparently does not flourish in areas of continual hot temperatures as indicated by its absence in all fully tropical climates. Pieris protodice differs from P. napi in its absence all along the immediate Pacific Coast from northern California through Alaska in keeping with its dislike for cool, cloudy areas. It is absent in the northern half of Alaska but eastward its distributional limits parallel those of the timbered area, as with P. napi. East of the Hudson Bay, however, P. protodice is absent north and east of Ottawa and New York State. The reasons for this event are baffling unless the present distribution of the species has been greatly disturbed by man in the eastern regions of North America as has been found true for Colias eurytheme. Pieris protodice prefers sunny, open fields as compared with P. napi which prefers partially shaded, damp areas. The cutting of the eastern deciduous forest to produce open fields for agriculture has aided both Colias eurytheme and Pieris protodice in extending their ranges and population density. It has probably a^so reduced the range and population density of Pieris napi since the habitats are conflicting. This probably accounts for the known diminution of Pieris napi in the region of Cambridge, Massachusetts where they were once exceed- ingy abundant, and now rare. 78 HOVANITZ /. Res. Lepid, FIGURE 2. Map showing the North American distribution of Pieris (Pontia) protodice. (0!73-^3, PIERIS DISTRIBUTION 79 Pieris protodice is found throughout the southeast from Long Island to the southern part of Florida and around the Gulf of Mexico to Guatemala. It is very abundant in the highlands of Mexico but may be found nearly everywhere except in woods or forested areas. It is found from the southern tip of Lower California northwards, having a great preference for open areas. It is possible that its great abundance in southern California coastal areas is a recent event, correlated with the change of the natural perennial grasslands to annual European grasses, mixed with Brassica nigra, the black mustard. It reaches its greatest abundance, however, in the desert regions where conditions of cool nights, hot days, full sunlight and periodic extensive abundance of food plants {Cleome, Brassica, Sisymbrium, etc.) allow the species to swarm in clouds. In such areas, the population abundance is usually limited by the periodicity of available water. In the mountain areas, the species is spread from the lowest valleys to the top of the highest mountains. The author has collected Pieris protodice at the top of Mt. Whitney, 14,500 feet elevation, the highest point in the United States south of Alaska, and in the bottom of Death Valley and Imperial Valley below sea level. The relationship of Pieris occidentalis to Pieris protodice is here left in doubt, but for the purposes of the map (Fig. 2), they are considered forms of one species. This point will be considered in more detail in a future paper. There is no doubt that at higher elevations and cooler temperatures, populations exist which have a blacker, fuller pattern on their wings than populations which exist at lower elevations at higher temperatures. The preponderance of the evidence suggests that these differences are not hereditarily controled, but that they are dependent entirely upon the environmental conditions. Within the species as a whole, as with P. napi, the darker indiyiduals exist in the more northern latitudes at cooler temperatures than the lighter ones. However, this relationship must be reconciled with local factors. In Colorado, Brown (1957) has shown that P. occidentalis appears to form definite populations at high elevations whereas P. protodice forms populations at low elevations. F. Rindge, however, in a personal com- munications has found both in considerable number in the same valley in Utah. These points cannot be reconciled at the present time. P. protodice when exposed to low temperatures develops a color pattern (more extensive melanin) which is similar to, or identical with, P, occidentalis. They could very likely be genetically the same and if they are genetically the same, they are not deserving of a scientific or Latin name of significance nomenclatorially. This point has not been proved, however. In the case of geographical variations, such deviations are considered subspecific unless proved otherwise. In the case of altitudinal or seasonal differences, such as in Pieris protodice or Colias eury theme, it is suggestd that Latin names not be applied except to geographically isolated populations and then in a subspecies sense only. 80 HOVANITZ /. Res. Lepid. FIGURE 3. Map showing the North American distribution of Pieris (Pontia) sisjmbrii. (i):73-h> 1962 PIERIS DISTRIBUTION 81 Pieris (Pontia) sisymbrii. (Fig. 3) The Pontia subgenus of the Pierids all prefer dry habitats with considerable sunshine exposure, unlike Pieris napi or even Pieris rapae which prefer or can survive in shade and high humidity. Pieris sisymbrii is one of these. Its habitats are always open, highly exposed, usually rocky places. It prefers cool temperatures but with high solar radiation. It lives in the ’spring time in, the desert, as in the Mohave Desert where it flies at temperatures of highs not over 70° F, even though the temperatures later in the year will be over 100° F. It also lives in the mountains at elevations of over 10,000 feet in June and July. In the coastal area of California, its habitats are usually dry, rocky areas (serpentine, etc.) which are more dry and more exposed to solar radiation than its general position on the map would indicate. Geographically (Fig. 3), the species exists throughout the moun- tainous areas of the west from the Yukon Territory to Mexico. It probably also exists in northern Mexico but there are no records. Outside of this area, there are a series of locations in the Northwest Territories from the Great Bear Lake southward past the Great Slave Lake to Lake Athabasca. Between these locations and the Rocky Mountains there are no locations known. This is probably due to lack of many suitable habitats. Collectors would do well to increase the known distribution of this species. It has a wide distribution but is characterized by the extreme isolated nature of them. Its habitats are narrowly restricted even on the desert and the adults have a short adult life, flying for only a few weeks during the year at any one place. The species aestivates and hibernates as a pupa which is very hard and impervious to water. There is relatively little variation between the individuals in a population in this species, and also relatively little variation geographi- cally between populations. Some variations have been described, how- ever, and will be considered in a subsequent paper. Pieris protodice is a species which is generally distributed and has many generations per year; it is therefore genetically and physiologi- cally pliable as is necessary to meet these changed conditions. On the other hand, Pieris sisymbrii is greatly restricted to isolated habitats, and the adults fly only a short time; thus, the species genetically may be said to have developed a narrower range of tolerance through a nar- rower range of selective factors of the environment. Pieris (Pontia) heckerii. (Fig. 4). Of all the species of North American Pieris, P, heckerii has the most restricted geographical range, extending from south-central British Columbia east of the Cascade Mountains southwards east of the Sierra Nevada into the deserts of southern California (Fig. 4). Its eastward limits coincide with the Rocky Mountain system in Montana, Wyoming and Colorado. The species does not extend into New Mexico so far as is known, unlike Pieris sisymbrii. The most southern-known limits are southwestern 82 HOVANITZ J. Res. tepid. FIGURE 4. Map showing the North American distribution of Pieris (Pontia) beckerii. (i):7}-Sl. 1962 PIERIS DISTRIBUTION 83 Colorado, central Arizona, and southern Nevada, except for California where the species may be found in the foothills of the mountain ranges especially on the desert side. However, the species has been found on the coastal side in the area of the Santa Susana Valley (Moorpark), San Fernando Valley, Santa Clara River Valley and on the immediate coast from San Diego and southwards. Pieris heckerii exists at elevations of sea level at the coast and some parts of the desert to 9000 feet in the mountain valleys. It has always been found in areas that are relatively dry or desert-like with high solar radiation. Its distributional range coincides almost perfectly with that of the semi-arid brushlands extending from the rain-shadow valleys of central British Columbia, throughout the Great Basin and inter-moun- tain valleys of the Rocky Mountains. The factors that limit its distribu- tion in southern Arizona, New Mexico and northern Mexico are unknown. Perhaps there is a lack of proper food plant. The distribu- tion of the species in southern California south of the Great Basin is correlated closely with the distribution of Isomeris arhorea, its preferred food plant. It seems likely, however, that some other plant may also satisfy for this purpose. There is little doubt that intensive searching for additional poptilations of Pieris heckerii in the range of Isomeris will extend the distribution some considerable distance, especially in the south Coast Ranges of California, the foothills of the Sierra Nevada and southward in Baja California. Unlike Pieris sisymhrii which has a single short generation per year, in southern California, Pieris heckerii has a succession of genera- tions with the bulk of the adult flight being correlated with the time of of the winter rains. However, in the mountainous regions north of the Mohave Desert, the species has only one known generation, in the early summer. There is little geographical, local population, or seasonal variation in Pieris heckerii. However, some races have been described and will be considered in a later paper. LITERATURE CITED BROWN, F. M., D. EFF AND B. ROTGER. 1957. Colorado butterflies. Proc. Denver Mus. Nat. Hist. Nos. 3-7:1-368. HOVANITZ, WILLIAM. 1950. The Biology of Colias butterflies. 1. The distribution of the North American species. Wasmann J. Biol. 8:49-75. J Journal of Research on the Lepidoptera 1(1) :85-88, 1962 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Copyright 1962 DID THE CATERPILLAR EXTERMINATE THE GIANT REPTILE? S. E. FLANDERS University of California Citrus Research Center and Agricultural Experiment Station, Riverside, California The abrupt disappearance of the giant reptile during, the Cretaceous period brought to an end a very successful plant-animal system. For 120 million years the reptile population had roamed over large areas of the earth, subsisting in the lush vegetation covering such areas. From the standpoint of the reptile s subsistence, plant life was unlimited. It was an apparently inexhaustible resource. The possibility that the great reptiles disappeared as a result of malnutrition and starvation has been considered and discarded as quite untenable. That highly evolved plant-feeding animals do not normally become limited in numbers by starvation is axiomatic. Nicholson (1954) pointed out that the major influence which prevents phytopha- gous animals from reducing the earth s vegetation to extre'me sparseness is their attack by natural enemies. Exhaustibility of plant life, an inherent characteristic of many plant and animal relations, is rarely realized, being usually precluded through the regulation of the animal population by its natural enemies ( Flanders 1959). Thus, the Cretaceous populations of the vegetarian reptiles, by serving as prey for populations of highly efficient predatory reptiles, were kept at levels well within the limits set by the amount of plant life. At the same time these predators served also to suppress the evolution of competitive types of animals of similar size such as the mammal. The geological record of the Cretaceous period, however, reveals a series of events which may have culminated in the temporary destruc- tion of vast am.ounts of plant life, a period in which the vegetarian reptiles may have been so devitalized that they either failed to reproduce or fell easy prey to their predatory relatives. The critical events which could have led up to the disappearance of the giant reptile were: (1) the replacement of the ancient fern- like plants by phanerogamic plants; and (2) the emergence of the Lepidoptera, an insect order which with few exceptions feeds only on such plants (Imms, 1930). The background against which these events should be visualized 85 86 FLANDERS /. Res. Lepid. is the world-wide dominance of insects both in species and in popula- tion densities, a dominance which existed throughout the Age of Reptiles and the Age of Mammals. The pre-Cretaceous insect fauna appears to have been quite similar in kind and in numbers to the post-Cretaceous insect fauna except for the lack of species which fed on flowering plants (Carpenter 1952). A number of such species were subsequently derived from the pre-Cretaceous insect orders of Coleop- tera, Diptera, Hemiptera, Hymenoptera and Orthoptera, species whose immediate progenitors had presumably already acquired their natural enemies. According to Imms (1931), it is practically certain that the Lepi- doptera is pre-Tertiary in origin, highly organized individuals being represented in Eocene rocks. The paucity of insects in Cretaceous rocks is attributed to the absence in that period of suitable fresh-water deposits in which specimens could have accumulated and become fossilized. Plant-feeding insects like other plant-feeding animals acquire natural enemies which serve to conserve the plant life upon which they subsist. The fact that the conservation of plant life is a function of the natural enemies of Lepidoptera has been demonstrated during the past sixty years by the importation and establishment in North America of the natural enemies of the caterpillar for otherwise the reduction in brown-tail moth, Nygmia phaeorrhoea (Donovan), the satin moth, Stilpnotia salicis (Linnaeus), and the larch caseberer, Coleophora laricella (Hubner) (Clausen 1956). During the past three years the writer, by using the grain moth Anagasta and its natural parasite, Exidechthis, has demonstrated in the laboratory the role of a natural enemy in the conservation of plant material. In a constant environment parasitization continuously conserved the plant material despite the presence of a persistent feeding population of Anagasta. In nature, therefore, a community of plant-feeding animals con- sisting, say, of a pentatomid bug, a grasshopper, a moth caterpillar, a vole, a rabbit, and an ungulate depending on the same resource, grass- land vegetation, can remain stabilized for a long period of years ( Elton 1946). Such stabilization, however, consists in the total amount of plant life conserved, not in the relative abundance of the plant species involved, an abundance determined by the phytophagous animals through their effect on the competitive capacities of the plant species (Wilson 1949). Devastations of plant life by caterpillars can occur today whenever the regulating action of their natural enemies is disrupted by environ- mental factors. Graham (1939) reported that during a ten-year period ending in 1920 an eruption of the spruce budworm, Choristoneura fumiferana (Clemens), was characterized by flights of moths so numerous that in the tree tops they had the apparance of a snow storm. During this eruption the caterpillars destroyed a volume of wood (i):85-88, 1962 CATERPILLAR AND REPTILE 87 sufficient to supply all of the pulp mills then operating for a period of forty years. The capacity of a lepidopteran in the absence of its natural enemies to reduce great areas of lush vegetation to extreme sparseness and to maintain it so thereafter was demonstrated in Australia when the caterpillars of Cactoblastis cactorum (Berg) imported from Argentina in 1925 destroyed within six years approximately 50 million acres of the prickly pear, Opuntia spp. Dodd (1929) in reporting on this devastation stated that great care had been taken not to set free any of the natural enemies of the catrpillar for otherwise the reduction in prickly pear would have been of minor proportions. It is evident that the plant-consuming capacity of a caterpillar population could equal that of a giant reptile population and that this capacity because of the caterpillar’s very short life cycle could be attained at a much greater rate. Supplementing this capacity was a power of survival much greater than that of the giant reptile, the minimum food requirements of the individual caterpillar being infinitesimal relative to that of the individual reptile. Nevertheless, there is no reason to believe that the caterpillar and the giant reptile could not have subsisted together on the same vegetation if the caterpillar, immediately upon its Cretaceous emergence, had been subject to a regulation by natural enemies as effective as was that of the vegetarian reptile by its predatory relatives. The theory that the caterpillar exterminated the giant reptile rests on the assumption that between the emergence of the new insect order and its accessions of regulative natural enemies there occurred a brief period in which the caterpillar population regulated the world’s supply of plant food, a condition conducive to great variations in abundance of plant life and of caterpillar life, a condition characterized by "feast and famine" in which the caterpillar was able to survive but not the great reptile. The inherent weakness of the reptile was an extraordinary need for an abundance of plant material. Only a few years of plant scarcity could have exterminated it. Hordes of caterpillars, in rapid consump- tion of fig and breadfruit, laurel and willow, oak and magnolia, could have so restricted the spatial distribution of the giant reptile that either its diet was inadequate or it was unable to avoid its natural enemies. The small size of today’s descendent reptiles, the vegetarian turtle, the predatory crocodile, the snake, and the lizard, is evidence of the giant reptile’s elimination by starvation and predation. The abrupt end of the Age of Reptiles during the Cretaceous period is ascribed to a newly emerged order of insects, the Lepidoptera, on the supposition that for a brief period it regulated the world’s supply of plant life at starvation levels for the dependent reptiles. Thus, the giant reptiles which had survived during eons characterized by great 88 FLANDERS /. Res. tepid. changes in climate, continental uplifts, and different diets may have been exterminated by the lowly caterpillar. LITERATURE CITED CARPENTER, F. M. 1952. Fossil insects. 1952 U.S.D.A. Year Book. pp. 14-19. CLAUSEN, C. P. 1956. Biological control of insect pests in the continental United States. U.S.D.A. Tech. Bull. No. 113, 151 pp. DODD, ALAN P. 1929. The progress of biological centrol of prickly pear in Australia. Commonwealth Prickly-Pear Board, Brisbane Qld. 44 pp. ELTON, CHARLES. 1946. Competition and the structure of ecological communities. Jour. Animal Ecol. 15: 54-68. FLANDERS, S. E. 1959. Biological control. Jour. Econ. Ent. 52: 784-5. GRAHAM, S. A. 1939. Principles of forest entomology. McGraw-Hill, N.Y. 410 pp. IMMS, A. D. 1930. A general textbook of entomology. E. P. Dutton & Co., Inc., N.Y. 703 pp. IMMS, A. D. 1931. Recent advances in entomology. P. Blakiston’s Son & Co., Inc., Philadelphia. 374 pp. NICHOLSON, A. J. 1954, An outline of the dynamics of animal popula- tions. Australian Jour. Zool. 2:9-65. WILSON, F. 1949. The entomological control of weeds. Int. U. Biol. Sci. Ser.B. 5:53-64. Journal of Research on the Lepidoptera 1(1) :89-93, 1962 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Copyright 1962 FURTHER EVIDENCE OE THE DISTRIBUTION OF SOME BOREAL LEPIDOPTERA IN THE SIERRA NEVADA' C. H. ERIKSEN Los Angeles State College, Los Angeles, California Except where roads penetrate the high country, little or only scanty information is abaiiable concerning the animal life of the high Sierra of California. This condition is extant because these mountains are high, rugged and largely accessible only by trail. The fact also that collecting gear must be packed in and out encourages few people to undertake such studies. Our knowledge of the Lepidoptera fauna of the Sierra stems largely from the work of Garth (1935) and Tilden (1959). Garth’s study of the Butterflies of Yosemite National Park is notable since he not only provides a list of species taken within the confines of the park, but relates them to the life-zones in which they are normally found to fly. Tilden has refined the latter concept in his Tioga Pass studies by listing associations of smaller scope than life-zones. While backpacking the John Muir TraiP during the summers of 1953, 1954 and 1955, the author made spot collections of Lepidoptera at various locations along the route (Figure 1). Such was a brief attempt to add information similar to that of Garth and Tilden to our knowledge of the day-flying Lepidoptera in high areas from Yosemite on the north to Mount Whitney on the south. All forms collected and altimdes at which they were taken are listed in Table 1. To alleviate any confusion, all nomenclature is after McDunnough (1938). Altitudes were determined from Starr ( 1953) . Since Garth (1935) and Tilden (1959) have shown that various species fly only within certain altitudinal ranges while others are unrestricted in their flight, there is little need to repeat similar fiindings here. However, the data collected does extend the known altitudinal range of several species. The following were taken at elevations higher than previously recorded and are associated with the life-zone of this extension. 1 My thanks to Nelson Baker, Santa Barbara Museum of Natural History, for help with identifica- tion of the material and W. Hovanitz for aid in preparation of this paper. 2The John Muir Trail follows a 215 mile route along the Sierra crest from Yosemite Valley on the north to Mount Whitney on the south. Except for a short distance out of Yosemite Valley (elevation 4,000'), altitudes range from approximately 7,000' to 14,I00'. 89 90 ERIKSEN /. Res. tepid. Argynnis mormonia (Arctic- Alpine) Euphydryas chalcedona ( Transition, Canadian ) Lycaena helloides ( Canadian, Hudsonian ) Plebeius aquilo podarce ( Arctic-Alpine) Plebeius saepiolus ( Arctic-Alpine ) Plebeius icarioides (Hudsonian, Arctic-Alpine) Plebeius acmon ( Hudsonian ) Garth does not list Lycaena helloides above the Transition zone nor Plebeius saepiolus above the Hudsonian. Even though Tilden does not actually say, he intimates that both species transcend all zones. In the event that any confusion may arise, both species are included in the above list. It is of interest that those forms listed in this paper, which also are found in Colorado, have already been collected in that state from life-zones here described as extensions (Brown et al, 1957). LITERATURE CITED BROWN, F. M., D. Eff and B. Rotger. 1957. Colorado Butterflies. Proc. Denver Mus. Nat. Hist., Nos. 3-7:1-368. GARTH, JOHN S. 1935. Butterflies of Yosemite National Park. Bull. So. Cal. Acad. Sci., 34(1) :l-39. McDUNNOUGH, J. 1938. Check List of the Lepidoptera of Canada and the United States of America: Part L Macrolepidoptera. Mem. So. Cal. Acad. Sci., 1:1-275. STARR, JR., WALTER A. 1953. Guide to the John Muir Trail and the High Sierra Region. Sierra Club, San Francisco, 130 pp. TILDEN, J. W. 1959. The Butterfly Associations of Tioga Pass. Wasmann J. Biol., 17(2) :249-271. I (i): 8^-93, 19^2 BOREAL LEPIDOPTERA 91 92 ERIKSEN /. Re$. Lepid. Arctic Alpine Mather Plateau 11,500* ; 7-30-53 ro €n Bighorn Plateau 11,500* 1 7-31-54 bH' ffH a- Tyndall Creek 11,000* i 7-29-54 pH cO - Upper Bubbs Cr, 11,000’ ; 7-29-54 sH Evolution Lake 10,990* s 7-26-53 a> Hudsonian Mdl* Palisade Low, L, 10,650* t 7-28-53 sH Cascade Valley 10,400* 1 7-5-54 Duck Uke Trail 10,100* s 7-5-54 - tH Purple Lake 10,000* S 7-6-54 CD ffH OI 1000-Island L. 9,850* 1 6-30-54 OI ■H (tH Little Pete Mdw, 9,100* s 7-12-55 O *H Shadow Creek 9,000* s 7-1-54 (O (H CM Shadow Creek 8,750* 1 7-2-54 Tuolumne Mdws, 8,700* j 6-29-54 CM pH Canadian Palisade Creek 8,125* 1 7-31-55 eH CO Quail Meadows 7,700* 1 7-8-54 CO iH tH .H ra Table 1. John Muir Trail Collection Sites, Listed by Elevation, and the Species and Number Taken Map Location Number 1 Papilio zelicaon Luc, Papilio rutulus Luc, Parnassius clodius Men, > -o m m Q X >. s, p « m O o Colias behrii Edw, Pieris sisynbrii Bdv, Pieris occidentalis Reak, Argynnis montivaga Behr > (0 •H G 0 1 o s m G G < Brenthis epithore Edw, c o •o 0 0 »H ns ^ • o » m X r@ $4 ^ * >, > 3 ^ U Q Euphydryas sierra Wgt, « m c •H a c >* o. 3 M Melitaea palla Bdv, Phyciodes laontana Behr Polygonia zephyrus Edw, (i); 8^-93. 1962 BOREAL LEPIDOPTERA 93 CO CM •H CM to CO <0 r-t pH CO CO - ■H •H - CM tn eH - pH •H CO •H - •H CM •H - rH u> - CM CM CO Nymphalis antiopa L. Junonia coenia Hbn. Basilarchia lorquini Bdv. Callophrys dumetorum Derulexa 3. & B. Lycaena edit ha Mead Lycaena helloides Bdv. lycaena cupreus Edw, Leptotes marina Reak. Plebeius aquilo podarce F. £ F. Plebeius saepiolus Bdv. Plebeius icarioides Bdv. Plebeius acmon West & Hew. Plebeius shasta comstocki _E22$ ^ ^ Philotes battoides Behr Lycaenopsis pseudargiolus echo Edw. Thoybes nevada Scud, Polites sabuleti tecumseh Grin. Celerio lineata Fabr. Pseudohazis hera Harr. Anarta cordigera Thun. 1(1) :95-96, 1962 Journal of Research on the Lepidoptera 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Copyright 1962 ARGYNNIS AND SPEYERIA BY WILLIAM HOVANITZ Department of Zoology ^ Los Angeles State College The use of names for nomenciatorial purposes involves judgment on two levels. The first level involves nomenciatorial laws, and the second is the judgement on the part of the taxonomist as to the most reasonable and useful application of a name for a biological category. When a name is being considered for use on the generic level, the first point is its nomenciatorial availability. Here, it is important that the name not be a homonym, and that it have priority according to the established laws of nomenclature. A species is designated as the type species of that genus. For example, the species paphia L. has been designated as the type species of the genus Argynnis. Once the type species has been established for a genus, and there are no doubts of the nomenciatorial status of either, the specific name or the generic name, no one who wishes to remain within established criteria for nomenciatorial uniformity and fairness has any right to change this name. The generic name is strictly applied, however, only to the one species. It is the prerogative of any individual to use the same generic name for other species if he deems that the species concerned ought to be considered congeneric. Strictly speaking, it is possible for each species in existence to be the type species for a different generic name. If each of these names were used, each species would be in a different genus and the beneficial effects of binomial nomenclature would be rendered worthless. The purpose of a binomial nomenclature is to indicate relationships by grouping related species into the same genus. How this should be done is the prerogative of the individual taxonomist. It is assumed that such a person will use the method that is best suited to his purposes in showing relationships. The genus Argynnis in the broad sense is composed of many species, some of which can be grouped into categories having biological similarities, but which are smaller in content than the genus when considered with a broad view. Generic names have been used for some of these smaller categories with various species designated as types. It is perfectly satisfactory from a nomenciatorial as well as a biological standpoint to do this. This is the situation with regard to the use of Speyeria. The type species of the name as a genus is idalia. 95 96 HOVANITZ /. Res. tepid. For those persons who wish to consider idalia as not congeneric with paphia, or with any other older name, Speyeria is a valid generic name. Or, it can be used in a subgeneric sense, in which case Argynnis would again be used as the generic name. This author believes that the broad use of the genus Argynnis is preferable to the practice of restricting the name to the paphia group of species, and further believes that the genus can be divided into some more or less satisfactory subgenera of which Speyeria is one. Our American species then would be designated Argynnis {Speyeria) idalia to show its relationship. The other American Argynnids then would also continue to use the name Argynnis rather than Speyeria. It is not the purpose of the editor of this journal to insist on any particular terminology for to do so would impinge on the freedom of the scientific worker, who alone has the right to make his choice. This editor can only insist on sound data to back the decisions of authors in their use of terminology, as in drawing conclusions from their work. advertisement TAKE CHLOROCRESOL ON YOUR NEXT TRIP Since the publication of the Chlorocresol method, field collecting techniques have been changed radi- cally. The Chlorocresol method eliminates the need for time consuming field mounting and/or relaxing of specimens. With Chlorocresol, specimens are re- tained in a relaxed condition for an unlimited period, allowing mounting to be done at a more convenient time or place. Field mounted specimens which would normally occupy a standard insect box may be packed in a container only 5" x 5" x IVi'- With specimens in a relaxed state, damage is minimized in the event of rough handling. Catalog No. 182 Chlorocresol, 50 grams $0.95 Chlorocresol, 100 grams 1.75 Chlorocresol, 200 grams 3 00 Styrene box, 5" x 5" x 114" per dozen 3.00 Complete instructions are included with each shipment. Prices are F.O.B. Santa Monica, California. California residents please add 4% sales tax. BIO METAL ASSOCIATES Box 61 Santa Monica, California Volume 1 Number 1 August^ 1962 IN THIS ISSUE Editorial 1 Variation in the silvering of Argynnis (Speyeria) callippe in the interior mountain area of south central California . . . Oscar Elton Sette 3 The effect of various food plants on survival and growth rate of Pieris .... William Hovanitz and Vincent C. S. Chang 21 General characteristics of the movements of Vanessa cardui (L.) J. W. Tilden 43 Three factors affecting larval choice of food plant William Hovanitz and Vincent C. S. Chang 51 The generic, specific and lower category names of the Nearctic butterflies. Part 1 — The genus Pieris . . Paddy McHenry 63 The distribution of the species of the genus Pieris in North America William Hovanitz 73 Did the caterpillar exterminate the giant reptile? .... S. E. Flanders 85 Further evidence of the distribution of some boreal Lepidoptera in the Sierra Nevada Clyde Eriksen 89 Argynnis and Speyeria William Hovanitz 94 FRONT COVER ILLUSTRATION PARALLEL COLOR VARIATION IN SOME CALIFORNIA NYMPHALIDS Column 1, E.chalcedona Column 2, E.editha Column 3, C.leanira Column 4, C. pal la Column 5, C.hoffmanm Details in Ecology 22: 259-284, 1941. Ci> E @F RilSEARCHJ THE LEFI Volume 1 Number 2 ]a^ary, 1963 Established in 1 962 Published at 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Edited by WILLIAM HOVANITZ WITH EMPHASIS ON ENVIRONMENTALLY AND GENE- TICALLY INDUCED VARIATION, population analysis, evolution, phylogenetic taxonomy, zoogeography, comparative morphology, ecol- ogy, geographical variation, speciation, physiology, etc. In short, qimlity ivork on any aspect of research on the Lepidoptera. THE PURPOSE OE THE JOURNAL is to combine in one source the work in this field for the aid of students of this group of insects in a way not at present available. The JOURNAL will attempt to publish primarily only critical and complete papers of an analytical nature, though there will be a limited section devoted to shorter papers and notes. RATES: $ 8.00 per volume, personal subscription. $12.00 per volume, institutional subscription. All amounts are in U.S. dollars, payable in the U.S.A. AUTHORS ARE REQUESTED to refer to the Journal as an example of the form to be used in preparing their manuscripts. Fifty separates will be supplied to authors free; reprints will be sold at printer’s rates if ordered at time galley proofs are returned. If proofs are not returned promptly, the editor reserves the right to withhold publication, or to proceed with no responsibility. All issues of the Journal will be copyrighted; in submitting a paper each author agrees that the material has not been published elsewhere. MANUSCRIPTS AND SUBSCRIPTIONS should be mailed to the address noted above. THE LEPIDOPTERA FOUNDATION To provide permanent security to the JOURNAL, THE LEPIDOP- TERA FOUNDATION is being set up. Memberships are now being accepted at TEN DOLLARS per year, eight dollars of which goes to your subscription and two dollars goes into a permanent capital fund, the earnings only of which may be used for current publishing expenses. In addition a scale of higher contributions is to be set up which will be announced later. All funds received in excess of the subscription rate will go into the permanent fund. LIFE MEMBERSHIPS are offered now at Two hundred fifty dollars; these include the Journal for life. INSTITUTIONAL AND COMMERCIAL DONATIONS AND MEM- BERSHIPS are also available; rates will be announced later, or on request. These contributions are deductible. HELP YOUR JOURNAL. Meetings of members may be held at times and places to be arranged. Journal of Research on the Lepidoptera 1(2) :97-108, 1963 1140 ’IF. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 196} COMPOSITION AND RELATIVE ABUNDANCE IN A TEMPERATE ZONE BUTTERFLY FALMA THOMAS C. EMMEL and JOHN F. EMMEL Reed College, Portland, Oregon md Stanford University, California The Conner Pass area, lying on the crest of the Sierra Nevada range in Placer County, California, provides a rich variety of habitats lying between 6900 and 8300 feet elevation. During the summers of 1956 and i960, the authors made an intensive study of this region and we have reported elsewhere (Emmel & Emmel, 1962a, 1962b) on the ecology and factors affecting distribution of the 74 species com- posing the butterfly fauna. We wish to report here the interesting data obtained on the faunal composition and relative abundance of species within the seven major Rhopalocera groups recorded in the Donner Pass area, and to evaluate the possible factors influencing this distribu- tion of species. INTRODUCTION Hovanitz (1958) has. recorded the occurrence of Rhopalocera families and genera at each five degrees of latitude and 1000 meters of elevation above sea level for the entire New World. However, this study did not include the numbers of species or genera in each family that occurred in each of these areas. Tilden (1959) divided the butter- flies of the Tioga Pass area under five plant associations; Emmel & Emmel (table 1, 1962a) partitioned the Donner Pass fauna under four plant associations. But these authors did not compare the faunas of each association or attempt to ascribe reasons for the greater number of species of a particular family in one association than another. The purpose of this paper is to report an examination of the composition and habitat associations of the Donner Pass fauna and from this examination to derive some understanding of the factors affecting the distribution and success of species in a butterfly family within a given set of environments. DESCRIPTION OF STUDY AREA & METHODS In a previous paper (Emmel & Emmel, 1962a) ^ four general vegetational associations representing the union of floras from three 97 98 EMMEL AND EMMEL /. Res. Lepid. life zones (Transition, Canadian, and Hudsonian) were delineated, and may be briefly reviewed. 1. The W et-M.eadow habitats (elevation 7000-7260 feet) support verdant expanses of grasses (Poa species) and sedges (Car ex), Lilium, Mimulus., Delphinium, Castilleia, and other wildflowers, with scattered thickets of willows (Salix) along the stream banks. Trifolium, Viola, Aster, Potentilla, and other lepidopterous food plants (see table IV) are common. 2. Dry-Meadoiu areas (elevation 6800-7600 feet) are moist in the spring but are dry throughout July and August. Fireweed (Epilo- bium angustifolium) , Achillia, Calyptridium, and Penstemon, as well as grasses, are typical plants. Food plants include Kibes, Polygonum, Lupinus, Gnaphalium, and others as listed in table IV. 3. Dense Forest (elevation 6800-7600 feet), mainly of Canadian- Zone trees, covers about one-half of the total Pass area. Red Fir (Abies magnifca), Pinus contorta, P. monticola, and Mountain Hem- lock (Tsuga mertensiana) represent the majority of the forest trees. Only three known butterfly food-plant genera are found here: Pinus, Arceuthobium, and Ceanothus, 4. The Montane areas (elevation 7160-8383 feet) are generally open talus slopes, where the dry, almost soilless environment discourages the growth of most trees. Included in the flora of these mountain peaks are Rabbit Brush (Chrysothamnus) , Sage Brush (Artemisia), Mule- Ears (Wyethia mollis), Dipinus, and Sedum species. Eriogonum, Astragalus, Potentilla, Ceanothus, and other food plants in Table IV, are widely distributed here. Thus the diversity of vegetational associations in this temperate- zone locality provides habitats and food plants for a considerable number of butterfly species, and it is of interest to compare the relative success of species in different families in colonizing this favorable yet rather small area (six square miles) in the Sierra. In 1956 and I960, the number of species (see Emmel & Emmel, 1962a, for a detailed listing) in each of the four vegetational zones was determined by intensive collecting throughout Donner Pass. In the course of studying the influence of meteorological conditions on the flight activity of these species in I960, daily counts of the numbers of flying butterflies were made within the 500 x 800 feet "Lodge Meadow” study area (see Emmel & Emmel, 1962b), which contained 46 species recorded for the Pass (see Table III). This area is composed of both wet and dry meadows. These counts were made by direct observation and represented numbers seen for each species between 10 and 12 a.m. daily. The total number of butterflies re- corded in this latter study was 7,720. To determine "success” in numbers of flying adults, the daily counts for numbers of individuals for each of the species in a family were added together, and this figure was calculated as a percentage of the total for all families. Thus the contribution of each family toward the composition of the Donner (2):97-!oS, 196} DONNER PASS BUTTERFLIES 99 Pass butterfly fauna can be compared on a percentage basis, both as regards numbers of species and numbers of individuals. OBSERVATIONS Some 74 butterfly species were found in the Donner Pass area, and these species belonged to seven of the major families of Rhopalocera found in the western United States. The number of species in each of these groups varied from 1. fof the Danaidae to 28 for the Lycaenidae, or in terms of percentage of the total Pass fauna, 1.4% to 37.8% of ail the species. Table I shows the composition of the Rhopalocera groups recorded in the total Pass area. Table II and Figure 1 show the composition of the butterfly fauna recorded in each of the four general habitats of the Donner Pass area; the figures in Table II represent the per cent of total species in each habitat that belong to each family, while the graphs in Figure 1 show actual number of species for each family in each habitat. It is imme- diately obvious that the montane and dry-meadow habitats have the largest faunas of the four areas, and that the Lycaenidae family is usually dominant as regards number of species. As stated earlier, it was possible to obtain some data on the number of individuals of each species (46 in all) occurring in the Lodge Meadow study area. The composition and relative abundance of the six major Rhopalocera groups recorded there are given in Table III. For the most part, there is a good correlation between number of species and number of individuals for each family. DISCUSSION AND CONCLUSIONS The number of butterfly species found in this six-square-mile area of the Sierra Nevada is unusually high. Tilden (1959) records only 43 species as occurring in the Tioga Pass region, representing a study area of approximately the same size although of somewhat higher elevations (over 9941 feet above sea level). Garth (1935) records about 100 species for all of Yosemite National Park (area 1,179 square miles, elevations from 2000 to 13,090 feet), and the present authors and Lloyd M. Martin (personal communication) can note from their experience that the average restricted Sierran locality seems to support around 35 to 65 species. Thus the Donner Pass area is particularly interesting from the standpoint of invesigating the reasons for the occurrence and abundance of a species in a specific habitat, for as noted earlier (Emmei & Emmel, 1962a) this region is a meeting place for a variety of habitats- and a high number of plant species, and these factors may be viewed as possible reasons for this variety of Rhopa- locera. The relative proportions in number of species among the Rhopa- locera groups involved seems to follow the relative number as found throughout North America north of Mexico (determined from Ehrlich, 1961 ). The descending order of these groups in terms of number of 100 EMMEL AND EMMEL /. Jl«, Lepid. TABLE I. Composition of the seven major Rhopalocera groups recorded in the Donner Pass area. Family No, of Species % of Total Species Papilionidae 5 6,8 Pieridae 8 10,8 Danaidae 1 1,4 Satyridae 2 2.7 Nymphalidae 20 27,0 Lycaenidae 28 37,8 Hesperiidae 10 13,5 TOTALS 74- 100,0% "In our earlier paper (Emmel 5 Emmel, 1962a) , 74 species and 2 "forms”, Anthocaris sara reakirti and Colias eurytheme amphidusa, were listed, making a total of 76 phenotypically-distinct entities DONNER PASS BUTTERFLIES 101 species is (both in the continent and local Pass faunas): Lycaenidae, Nymphalidae, Hesperiidae, Pieridae, Papilionidae (follows Satyridae in the continent fauna), Satyridae, Danaidae. However, when the total Pass fauna is divided in terms of occurrence in each of four habitats, patterns of apparent preference for certain habitats emerge — both for all families combined and for individual groups (Table II and Figure 1). The highest total number of species (60) occur in the montane habitats, the next highest (55) in the dry meadow areas, the third highest (44) in the wet meadows, and the lowest number (26) in the forest areas. The descending sequence of environmental support of number of butterfly species is thus: Montane > Dry Meadows > Wet Meadows > Forest. This pattern, obtained by comparing numbers of Rhopalocera species, is probably explained by the observed fact that the montane areas contained a greater variety of ' mico-habitats” than any other area. These micro-habitats included small stream areas, exposed mountain tops, open and brushly talus slopes, etc., ranging from 7100 to 8383 feet in elevation, the greatest range in altitudes for any of the four "macro- habitat” areas considered. Such variation in ecological conditions (especially from edaphic and climatic standpoints) permits a wide variety of potential food plants to flourish. Axelrod (I960) has commented on the marked diversity of habitats in montane areas and the possible influence of this condition on early angiosperm evo- lution through permitting rapid development and adaptive radiation. Such diversity also influences butterfly distribution, variation, and speciation (see Emmel & Emmel, 1962a; Le Gare & Hovanitz, 1951; Moeck, 1957), and provides potential areas for successful invasion by butterflies already occurring elsewhere. The large number of total species in the dry meadow areas is believed to be due to the great variety of known food plants occurring there; the wet meadows supported correspondingly fewer food-plant species. The jorest areas contained the least variety of micro-habitats and food plants; also, solar radiation, an important influence on butter- fly flight habits (see Emmel & Emmel, 1962b), was obviously at a considerably lower level in this habitat than in the other three. Table IV shows the number of known food-plant genera in each habitat for each Rhopalocera group. These and the preceding data agree with the theory that a greater variety in each of these Sierran butterfly groups is more likely to occur where micro-habitats and food plants are most varied, as these conditions, together with a high level of solar radiation, provide support for a wider selection of species than in such areas as pine forests. This theory can be applied equally to the question of why one group has more species than another group in a certain habitat. The Papilionidae are more successful in forest than in montane areas in terms of per cent of total species there, but they actually have the most species in the latter areas. So we must consider: 102 EMMEL AND EMMEL /. Res. Lepid. WET MEADOWS DRY MEADOWS FOREST MOr^TANE 4 PAPiLION/DAE 6 4 2 PIERIBAE DANAIDAE Is;:; SATYRIDAE HESPERUDAE Fig. 1. Numerical composition of the butterfly fauna recorded io each of four general habitats of the Donner Pass area. the NUMBER OF SPECIES i(2):^7-io8, 15163 DONNER PASS BUTTERFLIES 103 1) That habitat in which the group has its highest percentage of the total species population: Papilionidae Forest {1.1%) Pieridae Forest (15.4%) Danaidae Forest (3.8%) Satyridae Montane (3.3%) Nymphalidae Forest (38.4%) Lycaenidae Wet Meadows (45.5%) Hesperiidae Forest (19.2%) 2) That habitat in which number) : Papilionidae Pieridae Danaidae Satyridae Nymphalidae Lycaenidae Hesperiidae the group has the most species Montane (4) Dry Meadows, Montane ( 6 ) ( 1 in each habitat) Montane (2) Montane (16) Dry Meadows ( 22 ) Montane (10) ( direct The really significant consideration in evaluating a group’s contribu- tion to the fauna in a series of habitats would seem to be the number of species it has existing in each habitat. Of course, a further and perhaps more important consideration would be the number of indi- viduals of the group in that habitat, but these data were not obtained. However, speaking from the standpoint of number of species the butterfly groups are uniformly most successful in the montane and dry meadow habitats. As opposed to evaluating the most suitable habitat for most groups, one can also consider the success of individual groups in colonizing or existing in the various habitats. In the wet meadow, dry meadow, and montane habitats, the Lycaenidae are the most abundant group — • with 20 to 22 species in such areas. The Nymphalidae take second place (10 to 16 species) except in the forest zone where they lead the other groups in number of species. The Hesperiidae are third — with 5 to 10 species in each zone — except for being second in the forest areas (leading the Lycaenidae). The Pieridae are uniformly fourth (4 to 6 species) while the Papilionidae have only 2 to 4 species in each zone. There is only one species in the Danaidae (all habitats, however) and two in the Satyridae (montane habitat only). This apparent relative success of each group, in our opinion, is not due to a "better” adaptation (the result of natural selection) of a group’s species in one or another habitat, as Ehrlich (1962) has emphasized. Instead, from our observations, success is likely due to the fact that food plants of species in the seven Rhopalocera families were more varied (as to number of kinds; Table IV), generally more numerous (as to number of individuals of each plant species), and 104 EMMEL AND EMMEL /. Res. Lepid. TABLE II. Composition of the butterfly fauna recorded in each of the j. four general habitats of the Donner Pass area. Family Habitat Area““ Wet Meadows Dry Meadows Forest Montane Papilionidae 4.5 5.5 7.7 6.7 Pieridae 9.1 10.9 15.4 10,0 Danaidae 2.3 1.8 3.8 1.7 Satyridae 0.0 0.0 0.0 3,3 Nymphalidae 27.3 27.3 38.4 26.7 Lycaenidae 45.5 41.0 15.4 35.0 Hesperiidae 11,4 14,5 19.2 16.7 (1) TOTAL SPECIES 44 55 26 60 (2) PERCENTAGE OF TOTAL PASS FAUNA (74 species) REPRESENTED: 59.5% 74.3% 35.2% 81.1% “The numbers in each vertical column (file) represent percent of total species in that one habitat (the total number of species is found at the bottom of each column). The actual numbers of species of that family in each habitat, from which the preceding percentages were calculated, are shown in graphic form in Figure 1, ““See Emmel S Emmel (1962a) for map and complete description of all habitats in the Pass area. i(2);97-to8, 196} DONNER PASS BUTTERFLIES 105 more generally distributed for the Lycaenids than for the Nymphalids, for Nymphalids than for Hesperids, etc. It is concluded in this paper that the composition of a butterfly fauna is affected by the require- ments (or adaptation) of each butterfly species in that fauna for a certain food plant, in addition to a certain level of solar radiation and other factors as noted in Emmel & Emmel, 1962a. In turn, food plants obviously have grater chances of finding proper environmental condi- tions in a macro-habitat that contains a number of varied micro-habitats, each with particular edaphic and climatic conditions. The greater the number of host food plants, the greater the number of potentially successful butterfly species, and this theory is believed to account for the observed distribution of the 74 butterfly species (of seven major groups) among the four macro-habitats of the Donner Pass area. TABLE III, Composition and relative abundance of the six major Rhopalocera groups recorded in the Lodge Meadow habitat. Family No, of Species % of Total Species % of Total Individuals Recorded Papilionidae 2 4.4 4.7 Piaridae 4 8,7 9,8 Dana i da e 1 2.2 0.3 Nymphalidae 15 32,6 44.9 Lycaenidae 16 34.8 23,2 Hesperiidae 8 17.4 16,7 TOTALS 46 100.1% 99.6% SUMMARY 1. The Donner Pass area in Placer County, California, supports an extraordinary number (74) of Sierran Rhopalocera species, and its varied habitats and vegetational associations make the area par- ticularly interesting for investigating the basis for the occurrence and abundance of a species and a family in a specific habitat. 2. The number of species in the seven major families of the Pass area are proportionately equivalent to the numbers of species in these 106 EMMEL AND EMMEL /. Res, Lepid. TABLE IV, Number of known food plant genera in each habitat for each Rhopalocera family" Rhopalocera Habitat Area Family Wet Meadows Dry Meadows Forest Montane PAPILIONIDAE Salix — Ceanothus Cymopterus Salix (stream) Ceanothus Prunus Sedum FIERI DAE Cruciferae genera Cruciferae genera Pinus Cruciferae genera Trifolium Trifolium DANAIDAE - - - SATYRIDAE ( Grasses ) (Grasses) (Grasses) Grasses NYMPH ALI DAE Viola Castilleia Salix r’lantago Castilleia Aster SlrsEum Ribes Gnaphalium Lupinus Ceanothus Castilleia Aster Chrysopsis Ceanothus Salix (stream) Lupinus LYCAENIDAE Salix tupinus Ceanothus Lupinus Arceuthobium Pinus Ceanothus Lupinus trifolium Potentilla Polygonum Ribes Eriogonum Potentilla Calyptridium Polygonum Sedum Eriogonum Astragalus HESPERIIDAE Potentilla Grasses Sidalcea? Potentilla Grasses Sidalcea — Potentilla Grasses "Data summarized from text of Eramel 6 Emmel, 1962a. (2):97-Jo8, 1963 DONNER PASS BUTTERFLIES 107 groups as found for the whole of North America (north of Mexico). From most to least number of species: Lycaenidae, Nymphalidae, Hesperiidae, Pieridae, Papilionidae (follows Satyridae for North American fauna ) , Satyridae, Danaidae. 3. The highest total number of butterfly species in Donner Pass occurs in the montane habitats ( 60 species ) . The dry meadow habitats support 55 species, the ivet meadow habitats have 44 species, and the jorest areas have 26 species occurring in them. 4. The basis of this distribution is explained by the theory that a greater food-plant diversity occurs in the areas (such as the montane macro-habitats) that have a number of micro-habitats, each with particular edaphic and micro-climatic conditions, and that this host- plant diversity promotes the immigration and continued successful existence of more Rhopalocera species than in less diversified macro- habitats ( such as these Sierran forest habitats ) . The possible influence of such factors as solar radiation and climate acting directly on butter- flies rather than plants is also considered. 5. Data obtained on the number of flying individuals of all species found in the Lodge Meadow area showed there was usually a good correlation between number of species and number of individuals for each family in this restricted Sierran locality. SYSTEMATIC LIST OF RHOPALOCERA OCCURRING AT DONNER PASS, PLACER COUNTY, CALIFORNIA PAPILIONIDAE 1. Papilio zelicaon Lucas 2. Popilio indra indr a Reakirt 3. Papilio rutulus Lucas 4. Papilio eurymedon Lucas 5. Parnassius clodius baldur Edwards PIERIDAE 6. Neophasia menapia Felder & Felder 7. Pieris sisymbrii Boisduval 8. Pieris protodice Linnaeus 9. Pieris rapae Linnaeus 10. Euchloe creusa hy antis Edwards 11. Anthocaris sara form julia Edwards 12. Colias eurytheme Boisduval 13. Colias phUodice eriphyle Edwards DANAIDAE 14. Danaus plexippus Linnaeus SATYRIDAE 15. Ceononympha tullia calif ornica Westwood 16. Cercyonis sthenele oetus Boisduval NYMPHALIDAE 17. Speyeria cyhele leto Behr 18. Speyeria zerene zerene Boisduval 19. Speyeria coronis snyderi Skinner 20. Speyeria atlantis irene Boisduval 21. Speyeria mormonia arge Strecker 22. Boloria epithore Edwards 23. Chlosyne palla Boisduval 24. Chlosyne hoffmanni hoffmanni Behr 25. Phyciodes campestris montana Behr 26. Phyciodes mylitta Edwards 27. Polygonia zephyrus Edwards 28. Nymphalis calif ornica Boisduval 29. Nymphalis milberti Latreille 30. Nymphalis antiop a Linnaeus 31. Vanessa cardui Linnaeus 32. Vanessa atalanta Linnaeus 33. Vanessa virginiensis Drury 34. Vanessa car ye Hubner 35. Precis lavinia Cramer 36. Limenitis lorquini Boisduval & Leconte LYCAENIDAE 37. Satyrium calif ornica Edwards 38. Satyrium sylvinus Boisduval 39. Satyrium saepium Boisduval 40. Satyrium behrii Edwards 41. Satyrium fuliginosa Edwards 42. Strymon melinus Hubner 43. Callophrys johnsoni Skinner 44. Callophrys nelsoni Boisduval 108 EMMEL AND EMMEL /. Res. Lepid. 45. CaUophrys augustinus iroides Boisduval 46. CaUophrys eryphon Boisduval 47. CaUophrys dumetorum perplexa Barnes & Benjamin 48. Lycaena arota virginiensis Edwards 49. Lycaena edit ha Mead 50. Lycaena nivalis Boisduval 5 1 . Lycaena cupreus Edwards 52. Lycaena heteronea Boisduval 53. Everes comyntas amyntula Boisduval 54. Plebejus anna Edwards 55. Plebejus saepiolus Boisduval 56. Plebejus icarioides Boisduval 57. Plebejus shasta Edwards 58. Plebejus acmon Westwood & Hewitson 59. Plebejus {acmon?) lupini Boisduval 60. Agriades glandon podarce Felder & Felder 61. Glaucopsyche lygdamus hehrii Edwards 62. Philotes enoptes Boisduval 63. Philotes battoides intermedia Barnes & McDunnough 64. Celastrina argiolus echo Edwards HESPERIIDAE 65. Thorybes nevada Scudder 66. Pyrgus ruralis Boisduval 67. Pyrgus communis Grt. 68. Erynnis juvenalis Fabricus 69. Erynnis afranius Lint. 70. Hesperia juba Scudder 71. Hesperia nevada Scudder 72. Hesperia harpalus Edwards 73. Polites sonora Scudder 74. Polites sabuleti tecumseh Grin. ACKNOWLEDGMENTS The authors again express their appreciation to Mr. William N. Goodall of the National Audubon Society, Director of the former Audubon Camp of California at Donner Pass (Norden, California), for extending full coopera- tion to us during the collection of the data reported in this series of three papers on the Donner Pass butterfly fauna. Drs. Thomas Harvey and Kenneth Tanksley of the Camp staff kindly furnished plant identifications. Discussions with Mr. Lloyd M. Martin, Associate Curator of Entomology at the Los Angeles County Museum, have illuminated a number of points included in these papers. The authors wish to thank Dr. G. Frank Gwilliam, Associate Professor of Biology at Reed College, for his help in preparing the present paper. LITERATURE CITED AXELROD, DANIEL 1. I960. The evolution of flowering plants. In The Evolution of Life, Sol Tax, ed. Pp. 227-305. Univ. of Chicago Press. EHRLICH, PAUL R., & A. H. 1961. How to Know the Butterflies. Brown, Dubuque, Iowa. 262 pp. — — — - — ■, & RICHARD W. HELM. 1962. Patterns and populations. Science 137: 652-657. EMMEL, THOMAS C., & JOHN F. EMMELL. 1962a. Ecological studies of Rhopalocera in a High Sierran community — Donner Pass, California. I. Butterfly associations and distributional factors. Journ. Lepid. Soc. 16: 23-44. — — , 1962b. Ecological studies of Rhopalocera in a High Sierran community — Donner Pass, California. II. Meteorologic in- fluence on flight activity. Journ. Lepid. Soc., in press. GARTH, JOHN S. 1935. Butterflies of Yosemite National Park. Bull. So. Cal. Acad. Sci. 34: 37-75. HOVANITZ, WILLIAM. 1958. Distribution of butterflies in the New World. In Zoogeography. American Association for the Advancement of Science, Publication No. 51, Washington, D. C. Pp. 321-368. LE GARE, MARY JUDE, & WILLIAM HOVANITZ. 1951. Genetic and ecologic analyses of wild populations in Lepidoptera. 11. Color pattern variation in Melitaea chalcedona. Wasmann J. Biol. 9" 257-310. MOECK, ARTHUR H. 1957. Geographic variability in Speyeria. Mil- waukee Entomological Society. 48 pp. TILDEN, J. W. 1959. The butterfly associations of Tioga Pass. Wasmann J. Biol. 17: 249-271. Journal of Research on the Lepidoptera 1(2) : 109-1 13, 1963 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 1963 THE ARGYNNIS POPULATIONS OF THE SAND CREEK AREA, KLAMATH CO., OREGON Part I: The effect of the formation of Mt, Mazama on the area AND ITS POSSSIBLE INFLUENCE ON THE BUTTERFLY FAUNAS OF THE Sand Creek Basin. J. W, TILDEN San Jose State College, San Jose, California INTRODUCTION The populations of Argynnis (Speyeria) found in eastern Oregon share with similar populations in many parts of western United States, a tendency to great variation and to a breakdown of the distinctive appearance of many of the named geographical pheno- types (subspecies). The region drained by Sand Creek, lying east of Crater Lake in Klamath County, Oregon, is occupied by populations of Argynnis exhibiting this variation to a marked degree. The Sand Creek area was chosen for study not only because of the variability of its Argynnis fauna, but in particular because it is a region the recent geological history which is fairly well understood, and inferences may be drawn of the possible effect of this geological history on the several species of Argynnis now inhabiting this region. Part I of this paper will deal with the recent geological history of Mt. Mazama and the nearby lands to the east of Crater Lake, and the possible influence of these events on the faunas now living in this region. Part II will attempt an analysis of the variation of these Argynnis populations, and is postponed until much more field work can be done so that a proper appreciation of the extent of variation is possible. BACKGROUND The stability of certain isolated populations has been interpreted as due to long occupation of the present habitat. This time element has been construed as allowing stabilization, without undue influence from surrounding populations. Such conditions infer stability in the habitat, without the disturbing influence of earth changes in the area. Such stable habitats have been referred to as refugia. Opposed to such a condition of stability, many portions of the earth s surface have been subjected to changes more frequently, and some of these changes have occurred in very recent times. Eastern Oregon is a region of many lava flows, and scoria and cinder falls, and 109 110 TILDEN /. Res. Lepid. the effects of these actions are still clearly visible in numerous places. It is suggested that such phenomena must have had a major effect upon any butterfly populations existing in the regions where such events took place. THE SAND CREEK BASIN AND ITS RECENT GEOLOGICAL HISTORY Sand Creek rises at Anderson Spring, elevation 7088 ft, slightly northeast of Kerr Notch in Crater Lake National Park, and runs easterly, leaving the park in the near vicinity of the now unused East Entrance. It continues out of the park, the elevation gradually decreas- ing. The area chosen for study is that from Lost Creek Campground, Crater Lake National Park, 5972 ft., located between Sand Creek and its branch, Wheeler Creek, and the point at which Highway 232 ("Old 97”) crosses Sand Creek. This area is readily accessible. Work in the Crater Lake National Park was possible under a permit from the National Park Service. The Argynnis samples indicate that the following species are repre- sented: Argynnis {Speyeria) coronis Behr, zerene Bdv., callippe Bdv., egleis Behr, atlantis Edw., and hydaspe Bdv. Possibly others may be present in lesser numbers but have not yet been collected. It is inter- esting to note that Argynnis cybele {leto) appears to be absent from the region, and also from all of Crater Lake National Park, though common west of the Cascade divide, and also in the Blue Mountains of eastern Oregon. The scarcity in the Sand Creek Basin of suitable wet habitats, or the recency of earth surface changes, may be clues to its absence. Fair samples of these various species, collected by the author and also by others, have been examined, and it is evident that the degree of variation is surprising. As stated above, analysis of these and future samples is postponed until later, but enough information is now avail- able to make it apparent that the populations of each species involved are far from stable. Many individuals resemble no named geographic segregate. Some species are represented by individuals resembling two or more named segregates and it is evident that no equilibrium in phenotype has been reached by any of the several species inhabiting the area. Crater Lake is not a true crater, but a caldera, brought into its present form by the collapse inwardly of the summit of Mt. Mazama. In its most developed state, Mt. Mazama was more than 12,000 ft. in height. During the Pleistocene, it developed its greatest elevation, produced a great lava flow, and was subsequently glaciated extensively. The effects of this glaciation are visible around the present rim of the lake. Later, the present Crater Lake was formed by the collapse inwardly i(2):io^-Jij, 196} SAND CREEK ARGYNNIS 111 of the upper mile or so of Mt. Mazama, with attendant glowing avalanches of scoria and pumice and great clouds of pumice dust. The area east of the lake was covered with pumice and scoria from the glowing avalanches. The land still further east was covered by pumice falls that reached a depth of ten feet or more close to the mountain, and gradually thinned out easterly, extending in at least a very thin layer, approximately fifty miles to the east. It is believed that the glaciers, or some of them, must still have been present when the collapse took place, since the residue of scoria and pumice is light in certain areas around the rim. This would be the situation if the scoria and pumice fell onto glaciers and was carried away in part. If it had fallen in every case on bare ground, it would be left there. This may explain why the present pumice soils are more extensive in the non-glacial areas. It is believed also that the collapse took place in winter, because there is relatively little evidence of severe forest fires in the wake of the collapse. If the collapse had taken place in the summer, it has been reasoned that the forest fires surely would have been more extensive. There seems to be agreement that most or all of the vegetation was destroyed within the effective range of the scoria and pumice avalanches and of the heavier pumice fall. If this is correct^ the present vegetation of the entire region must have regrown since the collapse took place. The rate of recovery of the vegetation seems to have been inversely proportional to the depth of the pumice and scoria that covered a given area. The prevailing winds are believed to have been much the samx then as now, basically western, since the pumice was carried much further east than west. The Sand Creek Basin is situated so that it received very heavy pumice fall, and very probably, extensive to complete vegetational destruction. Carbon dating of wood that is imbedded in the pumice, and which is therefore assumed to have been charred by the glowing avalanches of pumice and scoria, sets the time of the collapse at 6640 ± 250 years. In round numbers, popular accounts state this as 6550 years. From this it is inferred that the Sand Creek Basin became a sterile habitat at that time. It is of interest to observe that such streams as Upper Sand Creek and Wheeler Creek, below Kerr Notch, have deep V-shaped canyons, geologically very young and without sign of glaciation such as that found on the rocks of the Rim above. From this it seems clear that Sand Creek acquired its present bed recently. Indeed, its present course may very well be consequent to the recent geological events of the area. The depth of the pumice is also clearly visible. Knowledge of the date when the vegetation was destroyed allows some speculation concerning present vegetational cover of the region. Recovery of the vegetation must have lagged for many years. The types of vegetation found in the total Crater Lake region permit some 112 TILDEN /. Res. Lepid. understanding of the relative times of recovery. In Crater Lake National Park, the forest along the southern boundary is a mature Sierran Coni- ferous Forest of Ponderosa Pine, Sugar Pine, White Fir, Douglas Fir and Incense Cedar, with an understory of Snow Brush {Ceanothus velutinus) and Greenleaf Manzanita {Arctostaphylos patula) . The forest of the flanks of the mountain on the southern and western slopes is a mature Mountain Hemlock-Shasta Fir Forest, with increasing amounts of Subalpine Fir at higher elevations. Such forests have relatively less understory. The highest elevations support Whitebark Pine. Northerly and easterly, in the areas of greater pumice fail, (as for instance the Pumice desert) the forest has not even yet succeeded in taking over the denuded and pumice-covered land. In the intermediate areas where the forest is slowly invading and colonizing the .pumice soils, the successful pioneer is Lodgepole Pine, which on such soils forms a singularly uniform and barren type of forest with very few underplants. The soils of the study area are formed of pumice and scoria and from the foregoing evidence, are derived from the pumice fall from the collapse of Mt. Mazama. The forest here is essentially Lodgepole Pine Forest, of the type found also on the pumice at higher elevations in the park. The soils and the forest are evidently contemporary in the Sand Creek Basin, with those occurring on pumice within the park itself. We should regard the forest cover of the Sand Creek Basin as equally recent. Dating of the pumice fall and the destruction of the vegetation give a reference point. Vegetational recovery must involve a very considerable period of time. This reasonably will be longer in areas where the pumice fall is deeper. While no exact figure can at present be suggested for the age of the current forest cover of the Sand Creek Basin, it must be very much less than 6640 ± 250 years. Very likely portions of it may be a fraction of this age. From this it seems evident that the Argynnis populations of the Sand Creek Basin (and similar areas) are occupying habitats that are the opposite of refugia. The time in occupation of contemporary populations on such disturbed areas is brief. DISCUSSION Assuming an area from which vegetation has been destroyed, with attendant soil changes, and with sufficient time elapsed to allow new vegetation to occupy the area, we may well expect certain things to take place, among which the folowing may be suggested. 1. The vegetation that recolonizes the land will in all possibility be different than that which previously grew in the area. Most likely many previous floral components would now find the area inhospitable. 2. Recolonization of the habitat by organisms would theoretically (2):i09-ii3, 196} SAND CREEK ARGYNNIS 113 be possible from any portion of the surrounding regions. It may be expressed by saying that recoionization could be expected to be muUi- dkectional unless some factor prevented recolonization from one or more directions. 3. Such recently recolonized areas would contain mixed populations the components of which may have moved in from any of the sur- rounding regions. 4. If such heterogeneous populations are studied before adequate time for stabilization has passed, it is predicted that great variation in individual phenotype will be present. Expressing this in terms more directly related to the Argynnis populations of the Sand Creek Basin, it is tentatively concluded that the variability of these several populations is in part a result of the recency with which they have established themselves in this area follow- ing a major change in habitat, and in part a result of recolonization by an unknown number of varying populations from the surrounding regions. Less tangible are two other possibilities. One is that the time during which the Sand Creek Basin was unsuitable for habitation by Argynnis was of sufficient duration to allow the surrounding populations which it separated, to diverge to some, extent. The other is that the recent forest cover of the pumice soils may differ sufficiently in its selective factors io that the eventual stable population may when developed differ from those around it. Thanks are due to the National Park Service, and in particular to Park Naturalist Bruce Black, for extensive courtesies and helpful information. REFERENCES BRIGGS, LYMAN J. 1962., When Mt. Mazama lost its top. National Geographic, 122 (1): 128-133. MOECK, ARTHUR H. 1957. Geographic variability in Speyeria. The Milwaukee Entomological Society, 48 pp. WILLIAMS, HOWELL. 1961. Crater Lake — The story of its origin. University of California Press, Berkeley & Los Angeles. XII + 98 pp., 10 pits., 8 figs. Journal of Research on the Lepidoptera 1(2) 1963 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 7563 CATERPILLAR VERSUS DINOSAUR? THEODORE H. EATON, JR. Museum of Natural History and Department of Zoology, University of Kansas A PAPER BY S. E. Flanders in the first issue of The Journal of Research on the Lepidoptera (Flanders, 1962) bears the intriguing title, "Did the caterpillar exterminate the giant reptile?” In his dis- cussion, the "abrupt end of the Age of Reptiles during the Cretaceous period is ascribed to a newly emerged order of insects, the Lepidoptera, on the supposition that for a brief period it regulated the world’s supply of plant life at starvation levels for the dependent reptiles.” The argument for this hypothesis rests in part on modern examples of large-scale devastation of certain species of plants by the larvae of Lepidoptera, and in part on assumptions concerning the time of origin and expansion of this order, the nature of plant life during the Cretaceous period, and the habits and relationships of the "giant reptile.” Many explanations have been proposed for the extinction of dino- saurs and several other groups of reptiles at the end of the Cretaceous, but there is not yet any general agreement as to the most important causes. The proposals are conjectures, having little or no direct evidence to support them. The theory under discussion adds another to the list. The consensus among paleontologists now is that one or two "causes” are not sufficient, and that the extinction of these animals is a more complicated phenomenon than it has seemed. While it is possible that destruction of food-plants by caterpillars has at times contributed to extinction of some herbivorous reptiles, and thus indirectly to that of their predators, the present writer thinks that a close look at what happened in the Mesozoic might be helpful in judging this hypothesis. Our knowledge of fossil Lepidoptera is' scanty, but the earliest known moth appears to be Loses triassica, described from Upper Triassic beds in Queensland (Tindale, 1945). It is known only from a fore and hind wing, of which the venation agrees broadly with that of Jugatae (Homoneura); Tindale placed it in a new, more primitive suborder, Eoneura. Thus the order must have originated more than a hundred million years before the extinction of dinosaurs, although leaf-eating caterpillars might not have characterized the earliest genera and families of moths. Reports of moths and butterflies from the Jurassic litho- graphic limestone in Bavaria are based on specimens now attributed 114 i(2):ii4-ii6, ip6j CATERPILLAR VS. DINOSAUR? 115 to other orders, especially Homoptera. Unless more information appears in Frank M. Carpenter s forthcoming volume on fossil insects in the "Treatise on Invertebrate Paleontology," we can say nothing specific about Lepidoptera of the Jurassic and Cretaceous; none are mentioned in the "Traite de Paleontologie,” vol. Ill (1953). In the Cenozoic, however, moths and butterflies of modern families are known, apparently from the Eocene on. It is reasonable to suppose that the rise of Lepidoptera coincided with that of flowering plants. These plants are known first from the Jurassic, but at least sixteen modern families are represented in Lower Cretaceous rocks; by the late Cretaceous many existing species of trees had appeared. The length of the Cretaceous was about 70 million years. From its close to the present time was a little less, say 63 million years, according to recent determinations (Kulp, 1961). But we cannot, in the nature of the case, show evidence that caterpillars appeared in great numbers in the late Cretaceous, or that they were then so free of para- sites or predators that they could, by worldwide devastation of all sorts of plants, bring starvation upon the reptiles. Some of the statements of Flanders concerning reptiles bear comment from the viewpoint of a paleontologist. For instance, "The inherent weakness of the reptile was an extraordinary need for an abundance of plant material . . . The small size of today’s descendent reptiles, the vegetarian turtle, the predatory crocodile, the snake, and the lizard, is evidence 'of the giant reptile’s elimination by starvation and predation.’’ A sauropod, hadrosaur, ankylosaur or ceratopsian dinosaur may have eaten a large quantity of vegetation, the amount depending on size and activity. These animals were present in the late Cretaceous and occupied a position comparable in some ways to that of the browsing mammals of the Cenozoic. But their extinction does not seem to have been abrupt; it was a slow, uneven decline probably extending through millions of years, and does not differ so far as we know from the decline and disappearance of various other groups of animals at earlier and later times. The flying pterosaurs, whose food was probably fish, declined and died out at about the same time as the dinosaurs. So also did mosasaurs, a specialized line of giant marine fish-eating lizards; so too the very different aquatic plesiosaurs, as well as a few late Cretaceous ichthyosaurs and carnivorous dinosaurs. No living reptiles are directly descended from any of the above- mentioned Mesozoic reptiles. Turtles and crocodiles originated in the Triassic and have continued with no fundamental changes to the present time. Lizards are known first from the Jurassic, snakes from the Creta- ceous, and their size cannot be taken as evidence of any changes among the giant reptiles. These giant reptiles were clearly not a single natural group, nor, on the other hand, were all members of the orders to which they belonged of great size. In the two orders of dinosaurs, for example, there were some species, both early and late, that failed to reach the size of an average alligator. 116 EATON /. Res. Leptd. Efforts to explain the extinction of dinosaurs and various other reptiles in the Cretaceous period ought to take into account a number of possible causes. The one proposed by Flanders does not give the answer to several problems, such as the disappearance of various un- related and ecologically different groups, not necessarily of large animals, and the survival beyond the Cretaceous of certain other reptiles, not necessarily small. There is no evidence of either a biological or a geological catastrophe of large proportions at the close of the Mesozoic, unless the slow retreat of epicontinental seas in some areas can be so described. Probably the answer will eventually be found in a combination of many factors. LITERATURE CITED FLANDERS, S. E. 1962. Did the caterpillar exterminate the giant reptile? J. Res. Lepid., 1 (1): 85-88. KULP, J. LAURENCE. 1961. Geologic time scale. Science, 133 (3459) : 1105-1114. TINDALE, N. B. 1945* Triassic inseas of Queensland. Rroc. Roy. Soc. Queensland, 56 (5): 37-46. Journal of Research on the Lepidoptera 1(2) :117-123, 1963 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 196} GEOGRAPHICAL DISTRIBUTION AND VARIATION OF THE GENUS ARGYNNIS I INTRODUCTION IL ARGYNNIS IDALIA WILLIAM HOVANITZ California Arboretum Foundation, Inc., Arcadia, California, and Los Angeles, State College, Los Angeles, California L INTRODUCTION The Genus Argynnis in the comprehensive sense covers a wide variety of form and size in Nymphalid butterflies. The geographical distribution of the genus as a whole is somewhat comparable to that of Colias but is actually wider in some areas due to a tropical group in the eastern hemisphere. A natural desire to have a greater similarity of type within a "genus” and a desire of some workers to "split” to the ultimate has left the genus in a diverse state of fragmentation. This has been aided by regional preferences and the difficulty for most Lepid- opterists to observe and study members of the group from continents other than their own. The members of the genus from the smaller isolated continents such as North America and South America have apparently either diverged slightly from the type on the largest land mass (Eurasia) or they have retained more primitive features. In North America only two of the several subdivisions of type within the Argynnis appear ever to have arrived for colonization. These two have diverged to such an extent on that continent that residents of that continent of natural inclination should consider them as of different genera {Boloria and Speyeria) . One of these {Speyeria) is only slightly different from one of the species groups of Eurasian Argynnis, namely, aglaja, adippe, niohe, etc. and placed by some in a genus Brenthis which is not the usage of Brenthis used in North America by others. The variety of the genus in Eurasia is much greater than in North America. It is natural to expect that the types present in Eurasia which would be most like those now present in North America would be those whose climatic adaptations would be the most similar to the conditions present in the area where a land bridge exists now or did exist in the immediate past. This is the area of Alaska and adjacent Siberia. The Boloria group is the best adapted to the climates of this area and may even now be considered for all practical purposes to have "continuity” across the barrier at the Bering straits. With the possi- bility of gene flow not very remote, it is not difficult to see why Boloria 117 118 HOVANITZ /. Res. Lepid. of North America and Asia (Europe) should be much alike. In fact, several species of Boloria can and do go under the same species name on both continents, just as do three species of Colias which have the same geographical relationships to the land bridge. The species which are the most remote in their climatic preferences from the climates of the Bering area are those species which are the most divergent on the two continental areas; these are the most southern types. Argynnis paphia is the member of a southern subgenus of Argynnis in Eurasia and this subgenus apparently never colonized America. Instead, a type similar to aglaja apparently did colonize America and from this presumably all the American Argynnids of the larger {Speyeria) type subsequently were derived. It is an hypo- thesis that the most extreme types in North America, such as idalia, diana and nokomts, which bear considerable pattern and color relation- ship to Eurasian species such as paphia, sagana, childreni, etc. were derived independently from them from aglaja — like ancestors under the selective influence of the comparable conditions of climate to which they were subjected (hot-humid summers). Comparative correlations of the facies of the butterflies with climates at different parts of the world will be made later in this series. The Argynnis of South America bear no close relationship apparent- ly to the Argynnis of any other part of the world. There is no easy way at the present rime for them to have arrived there by migration from other continents. Only two alternate proposals seem possible. The first of these is that these Argynnis were isolated many years ago (the early Tertiary) from the other Argynnis of the world and that they arrived there by migration from North America at a time when North America was inhabited by a more primitive type. They might also have arrived from Africa early in the evolution of the genus should there be reason to believe seriously in the past movement of South America away from a more close proximity to Africa. The second proposal on the origin of the South American Argynnis is one of independent origin in South America from some tropical or subtropical relative of the North American Argynnis. This proposal seems very weak in view of the lack of any close relative that would seem to be a likely ancestor. Dryas (Colaenis) , Dione and Euptoieta seem to be out of the question, though this may be only based upon superficial appearances. They feed upon similar plants, a factor greatly favoring the argument, especially with regard to Euptoieta which inhabits cold country in the Andes, feeds on violets and has much the same habits as Argynnis itself. The question comes up as to what is a genus. Should the entire group which has been pictured here be considered one genus, or should it be raised up to a sub-family or even a family. If the latter, then should the major groups be raised from subgeneric level to the generic.^ If so, then how many groups at this level should be recognized? The question is then condensed to one problem. There really is no biolo- gical entity which can be known as a genus. All hopes to the contrary, (2);ii7-I23, 1963 ARGYNNIS DISTRIBUTION 119 the genus is solely a matter of convenience in nomenclature, which should show as much as possible, phylogenetic relationships. The system should start at the bottom of the classification heirarchy. If it is known what should be considered species, sufficient related species ought to be put into a genus to make a reasonable group on the basis of numbers and morphological (genetical) similarity. Should too few species be put into a genus the advantages of the binomial nomen- clature are destroyed, in much the same way as would occur if every person had the family name Smith. A further criterion, in addition to size and uniformity, is geographical coverage. It is a natural phenome- non that isolation leads to differentiation. Related members of a species group may diverge slightly due to continental isolation and yet in all other ways be closely knit as a single group. The decision here must rest on the needs of the taxonomist in showing relationships and the usefulness to biologists. Nomenclature bears one major value: that of usefulness. If different generic names were to be used on each con- tinent, a degree of provincialism would develop which would be hard to penetrate; local butterfly books in North America and in Europe have now diverged to such an extent in terminology that the only way a novice can determine relationships between the butterflies of Europe and North America is to look at the specimens — -an ironic turn of events since that should be the purpose of nomenclature. As a result therefore, it is the hope that there may be a return to the larger genera of the past; this is the thesis upon which the use of Argynnis here is based, without necessarily accepting the assumption that the generic limits and the characters used for their delimitation should remain as in the past if more specific future work should show the need for change. IE ARGYNNIS (SPEYERIA) ID ALIA This representative of the North American Argynnis differs in appearance from the usual Speyeria probably more than any other except possibly for A. diana. The details of its pattern which are dis- tinctive and different from the standard Argynnis pattern are the greatly increased black pigment around the borders of the wings, even extending to the basal section and nearly covering the hind wings, but at the same time leaving free of black pigmentation, the marginal and submarginal rows of spots ( fig. 1 and 2 ) . The general facies therefore is distinctive and different from any other Argynnis. At the same time, the black pattern elements of the center of the fore wing are reduced rather than ex- panded as in all other related Argynnis. It therefore shows a develop- ment of a pattern not only unlike that of the Argynnis, including the subgenus Speyeria as indicated by Dos Passos and Grey, but a trend in a diametrically opposed direction. The distinctiveness of this species is such as to need no real description or detailed analysis. Likewise, from the study of specimens throughout the geographic range of the species, there are no geographical variations apparent, though there are in- 120 HOVANITZ /. Re$. Lepid. Fig. 1. Argynnis idalia, upper side; top: male; bottom: female, from Oak Park, Illinois, July 8, 1906. F. S. Daggett. i(2):ii7-i2h 19^3 ARGYNNIS DISTRIBUTION 121 ilfjflllliilililigF iisSllfiL iSisiSssSssilism immmmmmmmmmmm |bs::ss;s::l ^^asms: g^^SSSSSSSi rfw slow******* W ^Ms^sssi Fig. 2. Argynnis idalia, same as fig. 1 only lower side. 122 HOVANITZ /. Res, Lepid, SCALE 1000 FOOT CONTOUR Fig. 3. Map of North America showing the distribution of Argynnis idalia. (2):ii7-i2}, 1963 ARGYNNIS DISTRIBUTION 123 dividual differences in size and slight differences in the pattern observed on an individual basis. Perhaps statistical differences could, and will, be detected in the future. The distribution of this species is limited to a central and eastern location across the United States and a small part of Canada. The biological reasons for this restricted distributional range are not known. It can be seen from the map (fig. 3) that the species ranges from New England (specimens which I have seen come from as far as Massa- chusetts though records indicate Portland, Maine as the furthest point reached north and east), southward to North Carolina (probably erro- neous records indicate Georgia), and west to Colorado and North Dakota (probably erroneous records indicate Montana). There are records or indications in old literature for Arkansas but I have seen no specimens from there. The species seem to prefer the area south of the Canadian life zone, that is, the Transition life zone of sorts, though it is pretty hard to outline any definite climatic zonation. Records indicate that it is to be found in most of the southern counties of Michigan and the southern counties of Minnesota. Scudder (1889) writes: "This butterfly belongs to the Alleghanian fauna, though its distribution appears to be somewhat irregular. It inhabits lowlands and is much more abundant in the extreme eastern portion of its range than elsewhere, unless it be the western prairies.” Mention of it in Louisiana by Strecker is most probably in error. A further discussion of the range of this species will be deferred until the maps of other species have been published, which will be used for comparison. REFERENCES DOS PASSOS, C. F. AND L. P. GREY. 1947. Systematic catalogue of Speyeria (Lepidoptera, Nymphalidae) with designations of types and fixations of type localities. Amer. Mus. Novitiates No. 1370. — — — — 1945. A genitalic survey of Argynninae (Lepidoptera, Nymph- alidae). Amer. Mus. Novitates No. 1296. MACY, R. W. AND H. H. SHEPARD. 1941. Butterflies (of Minnesota). Univ. Minn., Minneapolis. MOORE, SHERMAN. I960. A revised annotated list of the butterflies of Michigan. Occasional Papers of the Museum of Zoology, University of Michigan. No. 617. SCUDDER, S. 1889. The Butterflies of the Eastern United States. Cambridge, Mass. Journal of Research on the Lepidoptera 1(2) :124-134, 1963 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 196} THE RELATION OF PIERIS VIRGINIENSIS EDW. TO PIERIS NAPI L. SPECIES FORMATION IN PIERIS? WILLIAM HOVANITZ California Arboretum Foundation, Inc., Arcadia, California, and Los Angeles State College, Los Angeles, California In a previous paper (Hovanitz, 1962), it was indicated that a fuller discussion of the relationship of Pieris napi L, to Pieris virgin- iensis Edw. would be forthcoming. In that paper, the map showing the distribution of P. napi included the distribution of P. virginiensis but a closer distinction between the two was not indicated because of the reduced size of the map included. A more detailed map showing the relationships between the two biological entities under discussion is included with this paper, together with a description of the known facts of the lines of separation of the two "species” or races, and a summary of the events that have most probably occurred in relation to the two groups up to the present time. On the map in the previous paper (Hovanitz 1962), the range of Pieris napi L. for the eastern United States included virginiensis as well as napi itself. The race of napi with which we are concerned here is known as oleracea. This race blends northwards and westwards into other races by imperceptible gradations. Pieris napi oleracea (known also by the summer brood names, aestiva Edw., hyemalis Edw. and acadica Edw.) may be considered the name for the race of napi east of 90° Longtitude and south of 50° N. latitude, more or less, with the exception of the area south of New York state (38° N). Other names can be found in McHenry (1962). The area south of New York state and east of 90° Longitude is occupied by the race or species virginiensis. In New York state and other points in the vicinity, or close to the southern limits of oleracea, there are populations which in some cases appear to be more closely related to virginiensis than to oleracea, and in other cases, are more closely related to oleracea with occasional individuals similar to virginiensis. The status of the line of contact between oleracea and virginiensis is not one of free and even blending such as might be expected of freely interbreeding individuals in a case of solely geographical isolation. Populations are rather clearly either oleracea or virginiensis as the case may be, with only occasional examples of intermediates giving some hint of the nature of introgression at the line of contact. 124 (2) -.124-1 34. 19^3 PIERIS NAPI RACES 125 The contact of these two species or races bears a close relationship to the similar situation between the species or races Basilarchia astyanax and B. arthemis in the same general area (Hovanitz 1949). Specimens of P. napi studied in various collections have permitted a map to be drawn showing the detailed distribution of oleracea and virginiensis (fig. 1). On this map, locations where oleracea (or napi in the species sense) are known are shown by the letter N and virginiensis locations are shown by the letter V. In addition to the indications on the map, there are reasons for assuming that each of these races overlaps the other even greater than seems indicated on the map. Due to lack of specific data (e.g. "New York”), some of these indications could not be shown on the map. These details will be mentioned below. On the map, it can be seen that oleracea extends southwards in the mountains of New England to Connecticut where it comes in contact with virginiensis or something like virginiensis . It is doubtful that these specimens are truly virginiensis, though they do bear some resemblance to them. Instead they are more likely abberant oleracea, which are genetically tending in the direction of virginiensis through introgression. In addition, the habitat change out of the Canadian forest typical of oleracea to the "Transition” lowlands might lead to a phenotypic alteration in facies. Populations of considerable size are definitely known at the present time in Vermont and New Hampshire. They are not now known in the vicinity of Boston, Mass, though they were common there before 1850. The following bears witness to a reduction in distribution of napi in the 1800 s: From Scudder (1889) : "This butterfly is a member of the Canadian fauna . . "It is found throughout New England although seldom abundant south of the annual isotherm of 48°. Northward and eastward it is everywhere abund- ant and it continues to be so as far south as Williamstown, Mass. (Scudder), Dublin, N. H. (Leonard) and Portland, Me. (Scudder, Verrill). South of 42° N. L. or the northern boundary of Connecticut, it rarely occurs, although it has been taken at Farmington, Conn. (Norton), Newport, R. I. (Miss Coggeshall teste Higginson) and Fire Island Beach, Long Island (Smith). Even in the north, however, there seems to be some local causes for its abundance.” "Mr. Lintner, writing in 1864, says that in Schoharie, N.Y., 'it was extremely rare until recently,’ and previous to 1857, 'he had taken it but once in three years . . "Mr. Bacon of Natick, Mass, says that the insect by no means disturbs cabbages and turnips as it did fifteen or twenty years ago.” "I recollect once seeing the college yard in Cambridge, I think it was about 1857, fairly swarming with P. oleracea. It is now never found, to my knowledge, anywhere in the region about Boston, and I think it is wholly confined to the less cultivated and especially the hilly distrias of New England.” FromKlots (1951): "Napi has . . . suffered a great restriaion in habitat, and probably in habits, for it is now almost entirely limited to shaded Canadian Zone forest . . "Not recorded south of the Catskill Mountains in New York”. 126 HOVANITZ /. Res. Lepid. From Klots (1935) of Edwards, "He mentions a female taken by Mead in the last week of June at Stony Clove in the Catskill Mountains . . . many eggs were obtained . . . from the chrysalis an oleracea emerged.” "Stony Clove is in excellent territory for napi {oleracea) but I know of no records of virginiensis from there.” "I have . . . seen specimens of virginiensis from Big Indian Valley in the Catskills, where the environment is more suited to virginiensis.” "In 1931, near McLean, Tompkins County, New York, I was able to obtain undoubtedly authentic eggs, larvae and chrysalis of virginiensis, through watching females ovipositing on Dentaria diphylla Michx.” From Fiske ( 1901 ) : "Pieris napi. once common enough to be injurious to cabbages and other cruciferous plants throughout New England, has since the introduction of rapae become one of our rarest butterflies. Its range is now restricted to the northern and mountainous portions, and even in the fortresses of the White Mountains it is a scarce insea.” The eighteenth century idea of a "species” contributed a lot to a general confusion of the nature of the relationship between oleracea and virginiensis. It was then, and too often now, the general idea that a species was represented by a specific morphological type, without regard to the variation which a population may have within itself. Some of the best collectors and breeders of Lepidoptera have therefore become confused, having the all pervading desire to designate some biological unit as either a "species”, subspecies or variety when the true relation- ship may not be possible with only these tools of nomenclature. For example, Edwards (1881) says: "Virginiensis . . . has become a true species, although unquestionably, in a higher latitude, it appears as an occasional aberration only of Oleracea.” Also, Klots (1951) says: "Until very recently, virginiensis has been confused with napi, but it is now known to be a distinct species of more southern (Transition Zone) distribution.” Edwards (1868-1872) says: "In the Kanawha district, it replaces Oleracea which is yet unknown there. It is not uncommon in the month of May, frequenting open woods rather than gardens, and in this respect differing in habit from the allied species. I have never met with it later than June, though Oleracea, in the Northern States is most abundant after that month and continues breeding till the early autumn frosts.” "I have received specimens of Virginiensis from Mr. Wm. Saunders of London, Canada and am informed by him that it is there a rare insect.” DIFFERENCES BETWEEN P. NAPI OLERACEA AND P. NAPI VIRGINIENSIS The primary visible differences between P. napi oleracea and P. napi virginiensis can be observed in figs. 2 and 3. In figure 2 is shown the male and female of the upper side of virginiensis to the left; at the same location in figure 3 is shown the under side. These specimens are (2):i24-n4, 196} PIERIS NAPI RACES 127 Fig. 1. Map showing the distribution in the eastern part of North America of the introgression area of Pieris napi races virginiensis and oleracea. from near Pittsburgh, which although toward the north end of the range of virginiensis are quite typical of the populations south through the Appalachian mountains. The pattern of the upper side female should be noted as quite typical of this race, and which is only rarely present in oleracea. Of critical distinction, the pattern of the under side of the wings should be noted. The veins are outlined in a tan pigmen- tation, not heavy but rather broad in expanse, in appearance unlike any other race of napi. P. napi oleracea is shown in figures 2 and 3 by the central two specimens; these are from southern New Hampshire which is near the 128 HOVANITZ J. Res. Lepid. « 2 § a 6 w ^ On ..U o § a M .. B Ph S H • m 0\ X § ^ a ^1 S ,3 3 V N S m ^ G\ G ^ P rn a ph‘ O G\ H 1 ^ dl ^ e§ ^ e« ^ « 5 W -4 0 aJ t- a 1 ^ fN E (2);i24-134> 19^3 PIERIS NAPI RACES 129 Same as fig. 1, but all lower side fades' 130 HOVANITZ /. Res. Lepid. southern part of the distributional range of this species. They are how- ever typical in appearance. Note the difference between the two races on both the upper and lower surfaces: virginiensis is more of chalky white on the upper side as compared with oleracea; the wings on the upper side tend to be veined with black near the borders and cell in oleracea which is replaced with a brownish ''rapae-ty^t' pattern in virginiensis; on the under side the veins are heavily lined with black in oleracea in a way distinctive from that of virginiensis; and there is a yellowish suffusion on the under side of the wings in oleracea not present in virginiensis. On the other hand, the appearance of oleracea changes during the seasons while that of virginiensis does not. There is no indication that virginiensis has more than the one generation per year, this occurring in early May whereas oleracea has a succession of generations starting in April and extending throughout the year until the cold weather of the autumn. During this extended period the weather conditions are quite different, and the effects of this on the appearance of oleracea are profound. The appearance of the adults emerged during the warmer part of the season are quite different. A specimen of the summer brood is shown in figures 2 and 3 at the upper right. Note the slightly increased size but more important the almost complete absence of any pigment other than white on the wing surfaces. These summer forms have been given the names hyemalis, acadica and aestiva by Edwards for the general area under consideration. As a racial name, the only one of any significance however is oleracea. Throughout the ranges of these races, there is fairly good uniformity in pattern and character- istics; however, it should be pointed out that within each of these races and especially oleracea there is a great amount of variation which tends to blend the two together. For example, a specimen at the bottom right in figures 2 and 3 is from a population of oleracea but shows the "rapae- type” pattern of virginiensis. The underside of the wings however makes clear that the specimen is oleracea. Areas of typical oleracea may be described as follows: The Can- adian forest areas ranging from northern Michigan across northern Ontario and Quebec into eastern Quebec, New Brunswick, Nova Scotia, Maine and the mountainous parts of the New England states as far as Massachusetts. South of these points and until the typical populations of virginiensis are reached in New York state, Ohio and Pennsylvania^ the populations apparently are in a state of uncertainty regarding a complete divergence from the virginiensis or the oleracea genome and all the things pertaining thereto, including in addition to pattern differences, the physiological alterations required for the more southern existence, the univoltine life cycle and the changed food plant preferences. These intermediate populations lie in a line ranging from Michigan, across southern Ontario to Montreal in Quebec and southward to the Catskill Mountains of New York. Some of these can be distinguished as oleracea with virginiensis tendencies and others as virginiensis with oleracea tendencies, though the former outrank the latter. Populations which (2);i24-134, 1963 PIERIS NAPI RACES 131 would be considered oleracea with virginiensis tendencies would be the following, with an indication of the extent of the tendency: Michigan: Emmett Co., north part of southern peninsula (slight) Ontario: region around London, Huntsville, Sydney Field station, Marmora, Trenton ( considerable ) — - — : Aylmer, Bells Corners, Brittania, South March, Merivale, Constance Bay, Ottawa West (slight) Quebec: Harrington Lake, Gatineau Park (slight) Massachusetts: Cummington, Whately, Franklin Co., Mt. Greylock Populations which would be considered virginiensis with oleracea tendencies would be: Quebec: He Perrot (Montreal) New York: the Adirondacks, the Catskills, Ilion, McLean Ontario: Hamilton Populations with little sign of intergradation on the northern end of the range of virginiensis'. Pennsylvania: Scranton, New Brighton, Pittsburgh, Forbes Forest, Charlevoi, Showville, Washington Co., Fayette Co., Lawrence Co. Ohio: Canton West Virginia: Forks of the Kanawha, Coalburgh HABITS OF THE RACES The 'rather distinct break between the geographical ranges of oleracea and virginiensis would seem to need some isolation barrier for the maintenance of its integrity. This barrier is clearly not one of water or mountain but rather one of habitat selection and the accompanying biological changes in the physiology of the races concerned. For some idea of the field habits of these races, we may refer to other authors: Klots ( 1951 ) says of virginiensis'. "Until very recently, virginiensis has been confused with napi, but it is now known to be a distinct species of more southern (Transition Zone) distribution. Its dark markings are a more smoky, diffuse brown and it lacks the yellowish tint beneath, on the HW and apex of the FW, which napi usually has. It is single brooded. More even than napi, it is limited to woods and very local.” "Food: — Toothwort {Dent aria diphylla) , probably other Dentaria. One brood. Adults in early May (central New York.” "Range:— Transition Zone, Ontario, central New England and New York, s. to Virginia (TL. Kanawha, West Virginia) He says of napi oleracea'. '^Napi has . . . suffered a great restriction in habitat, and probably in habits, for it is now almost entirely limited to shaded Canadian Zone forest ... A virginiensis lacks yellowish beneath on the HW and apex of the FW and has the scaling along the veins of the HW and costa and apex of FW beneath much more diffuse and powdery looking. A rare dark form of P. rapae lacks dark spots above, but never has the dark-lined veins of napi, and has heavier scaled, less fragile looking wings.” Food. Formerly, at least, nearly any cultivated or wild cruciferae; now probably chiefly limited to native, woodland species, such as toothworts (Dentaria), Rock cresses (Arabis), Winter and Water cresses (Barbarea), etc. 132 HOVANITZ J. Res. Lepid. Three broods. Adults in late April (Massachusetts).” "Not recorded s. of the Catskill Mountains in New York.” Clark (1951) says of virginiensis: "Range :---One record, Frederick County northwest of Cross Junaion on the Bloomery Road (Rank 698) about 200 feet east of the West Virginia line, April 24, 1938.” "Variation — This species varies considerably in the intensity of the bordering of the veins on the under side of the hind wings. This bordering is pale in ail the individuals from Virginia and adjacent West Virginia that we have seen, some having the under side of the hind wings almost immaculate white.” "Occurrence: — Confined to rich deciduous woods and extremely local, occurring in limited numbers at usually widely separated localities. The reduction in the numbers of Pieris virginiensis and its present occurrence only in widely separated localities are undoubtedly the result of deforestation, which has greatly reduced the areas in which it is possible for it to maintain itself.” "Pieris virginiensis keeps strictly to the woods.” "Season: — One brood. This species appears shortly after the middle of April and flies until about the middle of May.” Clench (1958) writes of the same: "Pieris virginiensis . . . common in woods . . . one brood: Late April to mid-May. One . . . capture ... on 4 June, 1958 is surely a freak occurrence . .” Scudder ( 1889) writes of the habits of oleracea: "Food plants. This caterpillar feeds on various cultivated cruciferous plants, such as turnip {Brassica rapa) , cabbage {B. oleracea), radish {Raphanus sativa) , horseradish {Nasturtium armoracia) and mustard {S inapis ) . Mr. Lintner considers that it prefers turnip to cabbage, for he once obtained fifty eggs fom a turnip bed and could find none on adjacent cabbage plants of the same age, and this has been my experience.” Hoy, on the other hand, thinks it is found mostly on mustard. But it also occurs on some of our native plants, such as Arabis drummondi (Couper) and A. perfoliata (Fletcher), and Fitch says it occurs abundantly on the water cress {Barbarea vulgaris)." "Life History. This butterfly, which appears to be generally triple brooded, passes the winter in the chrysalis, the first brood has been seen as early as April 18, but usually appears between April 27 and May 9, a week or there- abouts after P. rapae. For several years Fitch observed it at East Greenwich, N. Y. for the first time on May 8 and 9 and only once as early as May 6; in northern localities it is sometimes as late as the third week in May. It usually becomes abundant between the 10th and 15th of the month and continues unti the end of the first week in June. The eggs are laid during the last half of May and early in June, and hatch in from 5 to 10 days . .. . the second brood . . . makes its advent during the last days of June or early July . . ., this brood in oleracea seems never to be very abundant . . . and to maintain itself for a comparatively short time; it becomes common by the end of the first week in July and sometimes disappears . . . before the end of the month . . . "The third brood appears as early as the last days of July but usually not before August; it . . . lasts until early September . . . Occasional specimens, . . . disclose butterflies late in September or early in Oaober.” "The species must be triple brooded in the north.” Edwards (1868-1872) wmts oi virginiensis: "In the Kanawha district it replaces Oleracea which is yet unknown there. It is not uncommon in the month of May, frequenting open woods rather than gardens, and in this respect differing in habit from the allied species. I have never met with it later than June, though Oleracea in the (2):i24-i34, 196} PIERIS NAPI RACES 133 Northern states is most abundant after that month and continues breeding til the early autumn frosts.” "I have received specimens of Virginiensis from Mr. Wm. Saunders of London, Canada and am informed by him that it is there a rare insect.” These data would tend to confirm the evidence derived from the data of museum specimens that oleracea is indeed a muitivoltine race which breeds from April all through the year until the Autumn frosts. Our own personal experience on napi in California does not conflict with this opinion. Here napi {venosa and castoria) are continuously brooded when possible in relation to temperature and moisture condi- tions. They go into diapause when conditions of the food plant are un- favorable; this is usually in the summer when the food plant {Dentaria) dries up. In fact much of the first winter generation does likewise as Dentaria does not stay green long. This accounts for the summer or late spring generations being less abundant than the winter or early spring generations. A similar observation has been made of the generations of napi in New England. The food plant difference between the races oleracea and virginiensis is important in only one way, namely in the restriction of habitat of napi to a different ecological niche. Dentaria is a plant of the open woods and partial shade. The flight period of virginiensis must be adjusted to meet the proper conditions of moisture requirement for the plant as well as for the degree of light availability in the deciduous forest where early spring may allow more light in the forest than in summer. The restriction of virginiensis to the "Transition” life zone forest may be correlated with this light and temperature relationship. The northern forest is cooler all summer long and the temperature needs of oleracea can be met usually in the open areas as well as in the partial shade of the forest. This would account for the different flight habits of the two races. In California, napi ( venosa and castoria ) fly only in the cool canyons where there is partial shade even though there is abundant food available for them outside in the form of wild mustards. Instead they are restricted to Dentaria of the shaded woods, while Pieris rapae and P. protodice fly only a few yards away in the direct sunlight and feed upon mustards. The food plant of oleracea so often indicated in the literature is winter or water cress {Barb area) which is apparently restricted to the vicinity of strearqs. There is indication that in the Catskill Mountains (New York) there may be a population of virginiensis on Dentaria while sympatrically there are populations of oleracea on Barbarea. If this is so, an excellent study could be developed to show the nature of the racial or specific isolation of these two at a region where both maintain reasonably distinct populations in close contact. The fact that oleracea flies in April, virginiensis in May and oleracea again in June and July indicates another additive isolation factor which would help to maintain the species or racial distinctness. By way of conclusion, the data presented show rather conclusively that Pieris napi oleracea and Pieris napi virginiensis are not as com- 134 HOVANITZ /, Res. Lepid, pletely separated as one might wish "species” to be but that they are not either fully racial or subspecific in the sense that their populations blend into one another as by a continual gene flow. It is therefore a matter of personal opinion whether they should be considered species or races for nomenclatorial purposes. For these purposes it is impossible to stand on the middle ground as a choice must be made. I choose to use the racial designation for the reason that the groups are almost completely allopatric, that is, the populations are geographically separ- ated and do not occur in the same place with certainty without inter- breeding and complete fusion. Also, intermediates do occur with high frequency in intermediate geographical areas. There is work to be done in this regard on the ground in important areas, such as, the Catskill Mountains, He Perrot (Montreal) or any other areas where a potential close unity of the two appears. The past changes in the distribution of the species have been greatly exaggerated, especially with regard to any competition between P, rapae and P, napi. The habits of the two species are so distinct that no compe- tition can reasonably exist. Past records of P, napi feeding on garden or cultivated cruciferous plants are undoubtedly correct but these larvae came from adults flying out of adjacent woods to lay their eggs on any cruciferous plants available. With the destruction of the woods in the New England states, the habitat for napi disappeared and then of necessity napi disappeared. This was the situation for the low lands. In the mountains, napi still exists as it did in the past. REFERENCES CLARK, A. H. 1951. The butterflies of Virginia. Smithsonian Misc. Col- lections, 116(7) : 1-195. CLENCH, HARRY K. 1958. The butterflies of Powdermill Nature Reserve. Research Report #7. Powdermill Nature Reserve, Carnegie Museum, Pittsburgh, Pa. EDWARDS, W. H. 1868-1872. Butterflies of North America. Philadelphia. The American Entomological Society. FISKE, Wm. F. 1901. An annotated catalogue of the butterflies of New Hampshire. New Hampshire College, Durham, N. H. HOVANITZ, WILLIAM. 1949. Increased variability in populations follow- ing natural hybridization. In Genetics, Paleontology and Evolution, Princeton. — — — — . 1962. The distribution of the species of the genus Pieris in North America. /. Res. Lepid. 1 (1): 73-83. KLOTS, A. B. 1935. On the life history of Pieris virginiensis Edws. /. N.Y. Entom. Soc., 43: 139-142. . 1951. A field guide to the butterflies of North America, East of the Great Plains. Houghton Mifflin Co., Boston. McHENRY, PADDY. 1962. The generic, specific and lower category names of the Nearctic butterflies. Part 1. The Genus Pieris. /. Res. Lepid. 1(1): 63-71. SCUDDER, S. H. 1889. The butterflies of the eastern United States and Canada. Vols. I-III, Cambridge, Mass., author. Journal of Research on the Lepidoptera 1(2) :135-156, 1963 1140 W. Orange Grove Ave., Arcadia, California, US, A. © Copyright 1965 THE MALE GENITALIA OF SOME COLIAS SPECIES BJORN PETERSEN University, Lund, Sweden In most families of Lepidoptera the genitalia show such distinctive characteristics that it becomes, easy to separate even closely related species. In the family Pieridae conditions are different. Pieris mpi L., P. hfyoniae Ochs., and P. ergane Hbn. are not separable on the basis of their male genitalia, nor are F. rapae L. and P. manni Mayer (Lorkovic 1928, Drosihn 1933). Similar conditions are present in the genus Colias. Before going into details a general description of the male genitalia of Colias will be given (cf. fig. 1). The distal part of the Vlllth tergum is more or less slender, Kusnezov (1915) has called this part the super- uncus, Warren (1950) named it the false uncus. At the base of the superuncus the tergum is incompletely sclerotized laterally. The saccus is rounded, sometimes pointed on its sides. The vinculum is long and slender. The tegumen shows a narrow process dorsally, the pseudouncus (Kusnezov 1915). The aedeagus has a long ventral arm. According to Warren (Lc.) the claspers at their proximal end are restricted to a blunt point, attached to the vinculum. The dorsal edge is said to be drawn upwards, parallel to the vinculum, and the dorsal terminal extremity attached to the tegumen. These statements do not correspond to conditions found in Protocolias imperialis (PL I, fig. 1). In this species the proximal part of the claspers is rather similar to that of other Colias species. The distal part is protracted so that the clasper obtains a shape more or less similar to that of many other Pierids. It may therefore be concluded that the clasper -head with the terminal tooth is the dorsal part of the clasper. The short pseudouncus and the marked distal lobe of the clasper give to P. impefialis a rather isolated position. In some characteristics it comes closer to the genera Catopsilia and Anteos. In other respects, however, such as the structure of the clasper head, P. imperialis is similar to Colias, thereby showing the relationship between these genera (cf.Klots 1929 a and b). In the genus Colias the genitalia are rather variable within a species or subspecies, as will be shown below. At the same time the differences between some species are only slight. Therefore a biometrical approach seemed necessary as a complement to the general descriptions of the 135 136 PETERSEN J. Res. Lepid. genitalia. The modes of measuring and the results are given first as important for the full understanding of the descriptions. The breadth of the superuncus. The edges of the superuncus in dorsal view may a) converge distally b) run parallely or c) diverge slightly from a narrow part near the base ( cf . fig. 2 ) . The breadth was on diverging superunci measured at the narrowest basal point, on others at the corresponding place. Fig. 1. The male genitalia of a Colias. a superuncus b pseudouncus c clasperhead d tooth of clasperhead e uncus The size of a certain part of an individual may be influenced by its general size. The breadth of the superuncus has therefore been correlated with a measure of size: the wing length. All species with genitalia similar to those of the three species previously investigated: C. hecla, nastes and palaeno are included in fig. 3. Here the breadth of the superuncus is plotted against the wing length in a logarithmic scale. The populations investigated are parted in two groups, the first with a slender superuncus, the second with a broad one. The regression line of the populations with a slender superuncus is: log y 1= 0.75 log X + 0.68; (r = 0.72; 0.02 >£> 0.01). The species along this regression line are both orange with a hecla- pattern, yellow with a ^^c/^-pattern, and yellow with a nastes-patitm. The species in fig. 3 with a broad superuncus are either orange with a hecla-^RttQm or yellow with a nastes-pRttQm. These two types of species can be grouped along two regression lines. Within both groups the correlation is statistically significant. t(2):i}5-i56, 196} COLIAS GENITALIA 137 Orange species: log y=: 0.80 log X + 0.77; (£ = 0.71; 0.05 > P > 0.02). Yellow species: log y 1.17 log X + 0.30; (r = 0.81; 0.05 > P > 0.02). Fig. 2. Differently shaped superunci with semicircular, less sclerodzed areas near the base., a convergent b parallel c divergent The means of the two groups are, however, not significantly separated. An analysis of covariance gives 0.1 > ^ > 0.05. The data on which fig. 3 have been based are tabulated in Tables 1 and 2. Among the species shown in these tables C. nastes is of special interest. In Scandinavia and in North America but also in Altai and the Sayan Mountains the superuncus is slender. C. nastes of Siberia and the Amur Province is intermediate between nastes of the Altai and the Sayan Mountains and the closely related, allopatric C. montium from S. Kansu. This species forms a transition to C. cocandica from Ferghana, the Issykkul, and from the Tianshan. The superuncus-breadth of all Asiatic individuals of nastes is plotted against wing length in fig. 4, The populations are grouped along two regression lines, though the correlation in none of the groups is statistically significant ( cf . Table 1 ) . An analysis of covariance, however, shows that the means of the two samples (shown in the figure by squares) are significantly different (0.01 > P> 0.001). In the yellow series of forms there is thus no sharp limit between species with slender and broad superunci. Even within a single popula- tion, that from the Amur Province — the variation is so great that it includes superunci typical f(Jr nastes nastes and others typical for cocandica (cf. fig. 4). In the orange series of forms conditions are different as will be treated later. In fig. 5 the superuncus breadth of arctic C. nastes and of its sibling in the Alps, C. phicomone, is plotted against wing length. The differentiation has proceeded so far that hardly any overlap is present. The wing length and breadth of the superuncus in some other Colias species, less similar to those treated above, are shown in Table 3. In the three pairs at the top of the table the genitalia are so similar that 138 PETERSEN /• Lepid. KOU J L I I ! I ( 1 I 1.30 1.32 1.34 l36 1.38 1.40 1.42 7.44 7.46 7.48 Fig. 3. Superuncus breadth of some Colias populations plotted against wing length. Data from Tables 1 and 2. O Population from N. America or northern Eurasia □ Population from Central Asia or Central Europe 7\ C. palaeno and interior rod 3hoNe=hecla patterns open mark=orange color rod helow—nastes patterns filled mark=yellow color no constant differences could be found. Warren (1950) states some differences between C. australis and hyale, a statement which was not confirmed by the specimens investigated. Two of the species of Table 3, C. cunninghami and P. imperialis, have an extremely broad superuncus. C. vautieri, which otherwise is 1(2): 135^156, 1.90 ^ .?«3 COLIAS GENITALIA log superuncus breath 1.85 O 139 1.80 1.60 o + log wing length U. I -J_ — __ 1.32 1.34 1.56 1.38 1.40 1.42 1.44 Fig. 4. Superanciis breadth of some Asiatic populations of C. nastes plotted against wing length. The lower square represents the mean of the dots, the upper square the mean of the remaining marks. • Altai and Sayan Mts. A Transbaikal O Siberia, Taiga zone + Amur Province rather aberrant, C. erate, and myrmidone fit well into the group of species with a broad superuncus, listed in Table 2. The remaining four species have a more slender superuncus though C. hyale and australis, like C. montium occupy a somewhat intermediate position. The form of the superuncus in lateral view-. During copulation PETERSEN /. Res. Lepid, 140 Log superurunijS breadth 100^ 95- o o o o 90- 85- 80- O O O O o oo o o o o o o o 75^ 65- 60- 55- 126 1.50 tsz 134- 1.56 log wing Length 138 UO 142 Fig. 5 Superuncus breadth of C. phicomone (open circles) and C. nastes (filled circles = Scandinavia, + North America) plotted against wing length. (2):i}5~i56, 1963 COLIAS GENITALIA .141 the superuncus together with the uncus is pressed against the body of the female, just as the uncus in the genus Pieris according to Lorkovic (1947). To make possible the bending down of the superuncus during copulation a semicircular area laterally on the Vlllth tergit is less sclerotized than the rest ( cf . fig. 2). This bending can be stated without studying any copulating pairs. In some specimens the superuncus is broken (cf. fig. 6), in some in addition kept between the two claspers. Fig. 6. Colias intenor with superuncus broken during copulation. The breadth of the superuncus of such specimens sometimes cannot be measured without maceration. A broken superuncus was found in all species where more than a few specimens were investigated, as shown in Table 4. Only in very few species the superuncus is a straight process protruding from the Vlllth tergit. Also when unbroken the superuncus usually forms a bow downwards. This bow is present, even in its most pronounced form, in animals which have not copulated, as, for instance, in all the thirty C. hecia specimens investigated which in 1952 were caught at the beginning of the flying time on the northern side of the Lake Tome Trask. To obtain a quantitative estimation of the variation of the form of the superuncus in lateral view, the angle of the distal part of the superuncus to the dorsal edge of the Vlllth tergum was measured. The result of this investigation is shown in Table 4. In general a broad superuncus is only slightly bent downwards while a narrow is bent more. The exceptions are rather few: C. interior, montium, and phicomone. 142 PETERSEN /. Res. Leptd. Species n log X M, log y r P b A. Orange species C, hecla, Scandinavia 51 1.331 1.70 -0.18 0.3-0. 2 -0.74 , Siberia, Dudinska 4 1.355 1.67 , Baffin, Isl. 2 1.347 1.79 , Greenland 3 1.362 1.74 C, hyperborea, N. Siberia 4 1,69 C. eurytheme, N. America 6 1.456 1.76 , slightly orange Texas, Houston 6 1.313 1.61 B. Yellow species a. hecla pattern C. palaeno, Europe 41 1.411 1.71 0.18 0.3-0. 2 0.50 C. palaeno, Siberia 9 1.377 1.71 0.57 0. 2-0.1 1.73 C. interior 2 1.311 1.67 b. nastes pattern C, nastes, Scandinavia 42 1.344 1.68 -0.09 0.6-0. 5 -0.24 , Siberia, Amur area 29 1.391 1.75 0.32 0.1-0.05 0.87 , Altai, Sayan Mts. 12 1.379 1.67 0.35 0.3-0. 2 0.86 , N, America 2 1.311 1.68 C. philodice, N. America 11 1.419 1.74 C. montium, S, Kansu 9 1.370 1.77 Table 1. Wing length (s<), in mm, and breadth of superuncus (y), in an arbitrary scale. and their correlation in some Colias species with a slender superuncus. Logarithmic scale. C. interior is in this character well separated from its allopatric sibling, C. palaeno, in having a rather straight but slender superuncus. C. montium has a rather slender superuncus which is straight as is that of cocandica. In this respect another of the southern species, C, phicomone, is intermediate between nastes and cocandica. The superunus of phicomone is, however, rounded and not more or less pointed as it is in all other species with a slender superuncus except C. interior, aurorina, and sagartia. By means of the two characteristics, i.e. the breadth and the form of the superuncus, it is possible to separate the two species C. viluiensis Men. and hyperborea Gr. Gr. which live sympatrically in N.E. Siberia. Of viluiensis 7 specimens from the Verchojansk area, Lutsha near Yakutsk, Vilutsk and Vilui have been investigated. The four specimens of hyperborea were from Sib. poL, the Lena Valley I (2) ; I )5-i56, 7963 COLIAS GENITALIA 143 Species n ”log X M, log y Species n '^log X M log y Orange species, hecla-pattern Yellow species, nastes-pattern C, chrysotheme 5 1.350 1.89 C. croceus, Europe 10 1.403 1.90 C. alpherahyi 1 1.436 2.02 , Asia 11 1.429 1.89 C. christophi 1 1.369 1.86 C. eogene 8 1.368 1.87 C. heos 12 1.474 1.99 C. cocandica 2 1.326 1.85 C. remanovi 2 1.446 1.91 C. phicomone 24 1.374 1.88 C. standingeri 6 1.395 1.83 C. sieversi 1 1.405 1.93 C. thisoa 2 1.417 1.90 C. sifanica 1 1.326 1.90 C. wiscotti 9 1.431 1.93 C. viluensis 7 — 1.90 Table 2. Wing length (x), in mm, and breadth of superuncus (y), in an arbitrary scale, ih some Colias species with a broad superuncus. Logarithmic scale. and Sredne Kolymsk.. The latter specimens all have a slender and strongly bent superuncus of the hecla-type, while all specimens of viluiensis have a straighter and broader superuncus. There is no overlap in any of the two characters; in the "angle-character” the gap is very wide. In Colias nastes from Asia there Is no similar correlation between the breadth and the shape of the superuncus. Among the 41 specimens investigated the coefficient of correlation is -j- 0.093 which is far below significance ( 0.8 > P > 0.7 ) . Conditions are thus quite different in northern Asia as regards the nastes- and hecla-stn^s of forms. In the nastes-SQnts dines including characters of the genitalia reach the Amur area, Transbaikal, the Sayan Mts, and the Altai, and forms intermediate between arctic and Central Asiatic ones are present in North Western China and in the mountains of Central Europe. In the orange series of forms an overlap of a northern and a southern species is present between 65° -68° n. latitude and no intermediates have yet been found between these species. Number of teeth near apex of aedeagus. In many groups of insects the number, shape, and position of the aedeagal teeth serve as good evidence to distinguish species. In the Colias species all the teeth are small, of a rather similar shape, situated near the apex. Variation in number is strong even within subspecies when compared with the differences occurring between species. Thus this character is without signifiance for the determination of individuals. A closer investigation of a material belonging to the group of species with a slender super- 144 PETERSEN /. Res. Lcpid. C. aurorina n M, log X M, log y 6 1,449 1.80 C, sagartla 5 1.437 1.76 C. australis 8 1.384 1.74 C. hyale 27 1.373 1.79 C . erate 13 1.387 1.94 C, myrmidone 7 1.425 1.95 C. cunninghami 2 1.367 2,01 C. vautieri 2 1.327 1.03 P. imperialis 1 1.356 2.05 Table 3. Wing length (x), in mm, and breadth of superuncus (y), in an arbitrary scale, in some Colias species. uncus, however, revealed some slight specific differences in the average number of teeth (cf. Table 5). The results of this investigation are; C. nastes has on an average fewer teeth than C. hecla from the same locality (Scandinavian material t = 2.45; 0.02 > P > 0.01, American material not differing significantly. There is a certain tendency of parallelism, both species having a lower number of teeth in Scandi- navia than in North America. This tendency is not significant in any of the species, however. C. palaeno and interior both have a low number of teeth, while the number is fairly high in C. meadi, philodice, and eurytheme. SPECIAL PART The genus Colias may be parted in two genera: Protocolias (type imperialis Btlr.) and Colias, the latter in turn in two subgenera Mesocolias (type vautieri Guer.) and Colias. The descriptions of the genera, the subgenera and their various species may be given most easily in the form of a key. 1. Small or medium-sized, orange butterflies with broad (0.40-0.45 mm) or very broad (0.5 mm), straight superunci. Superficial scales — if present— -yellow and black, broad and flattened in the distal part. Apex of aedeogus without teeth. Genus Protocolias and subgenus Mesocolias 2. r. Small to big, yellow, greenish or orange species with broad to slender (0.32 mm) superunci. Superficial scales in the black mar- gin—if present— yellow, pointed, hairlike or broader. A number of teeth at apex of aedeagus Subgenus Colias 3. i(2):i}5-i56, 196} COLIAS GENITALIA 145 2. Superuncus very broad and triangular. Pseudouncus short. Clasper- head with elongated tooth. Middle part of the clasper with a marked prong. Ventral arm of aedeagus broader in the distal end. Black and yellow suoerficial scales present. Pfotocolias imperialis Btlr. (PI. 1:1) 2’. Superuncus broad. Tooth of clasperhead slightly bent upwards. Ventral lobe (v.l.) of inner side of the clasper (Klots 1929 a) more marked than in any other species of the genus. Distal part of ventral arm of aedeagus showing charaaeristic shape. Superficial scales absent. Colias (Mesocolias) vautieri Guh. (PI. 1:2,3) 2”. Superuncus very broad, genitalia small, claspers short ( 1.0-1. 1 mm) but broad (0.6 mm). Distal part of ventral arm of aedeagus not broader than the proximal one. Superficial scales absent. Colias {Mesocolias) cunninghami 'Qth. (pi. 1:4). 3. Claspers caudally with a marked prong directed medially. Super- uncus straight and fairly slender. C. hyale L. (pi. 1:5) and australis Ver. 3’. Claspers without any marked prong ........4. 4. Claspers pointed in the middle part of the caudal edge (most easily seen from behind) ........5. 4’. Claspers not pointed in the middle part of the caudal edge 6. 5. Superuncus straight, on an average 0.40-0.45 mm broad C. erate Esp., C. myrmidone Esp. (PI. 1:6), 5’. Superuncus usually markedly bent downwards, 0.30-0.35 mm broad. C, sagartia Led. ( PI. 1 : 7 ) C. aurorina H. Sch. (PI. 1:8). 6. Superuncus on an average less than 0.32 mm broad (cf. text fig. 3) usually strongly bent downwards (cf. Table 4) .........7. 6’. Superuncus on an average broader than 0.35 mm, usually straight 13. 7. Superuncus only slightly bent downwards ..........8, 7’. Superuncus strongly bent downwards 9. 8. Superuncus in caudal view blunt-ended as in species with a broad superuncus. Inner side of -clasper of hecla-type (cf. PI. 11:34) C. montium Oberth. (PI. 1:9, 10). 8’. Superuncus in caudal view pointed. Inner side of clasper with well developed ventral lobe and ridged from its dorsal edge dorso- caudally towards the caudal part of the clasper (Fig. 14, 15) C. interior Scudd. (PI. I: 11-13, II: 14, 15), 9. Ridge from dorsal edge of ventral lobe towards caudal part of the clasper present as in C. interior. Ventral lobe less well developed than in C. interior (cf. Fig. 16). Caudal edge of clasper often strongly bent inwards ( most easily seen in caudal view ) . C. palaeno L. (PI. 11:16-25). C. Christina (PI. 11:26-28). C. palaeno: is most variable in all charaaers investigated except the breadth of superuncus (cf. the figures). One specimen from Abisko (42 specimens studied) completely lacks the ridge on the inner side of the clasper. Hence it cannot with certainty be separ- ated from C. nastes or hecla. The limited material of C. Christina appeared rather close to some specimens of palaeno, though it might be possible to separate the two species after examining a greater material. 9’. The caudal edge of the valva less strongly bent inwards ventrally. Inner side of clasper reminding of hecla-type (cf. fig. 34) ........10. 10. Clasoer (in lateral view) broadest rather ventrally 11. 10’. Clasper broadest in the middle part 12. 11. Aedeagus on an average with a higher number of teeth at the distal end ........................................C hecla Lef. (PI. 11:29-39). 146 PETERSEN /. Res. Lcpid. Frequency of superuncus broken 7/50 1 14/56 1 1 14/56 1 14/56 1/12 4/25 Ipdl 1 1 1 2/9 1 2/12 <D DO C < 82 s s cr> ID s s <T> CD s O to r- .It CO El CM =t- o - 95-104 1 CM - c. - *— J if - - CO CO - J- - - 3 - - *— 1 - •— J - CD - f—{ CO d- S - - - - - - ID - - it *— 1 CM - - - 3 m a- - - rH fH o c - CM #— J - - CO - 25-341 tD - - CN 1 •H CM •— 1 it •H Species, Population Popu of Table 1. C. hecla, Scand. Mts. " , N. America C. hyperborea C. meadi j C. eurytheme | C. palaeno, Scand. Mts. j ” , Europe | 1 " , Asia " , N, America C. interior [christina] m ■p 'O c o CD (0 «D CO m c m CO < " , N. America 1 C. philodice 1 C. beryl la Pop. of Table 2. C. chrysotheme i C. croceus, Europe 0} c 0) DO 0 o IT. Aedeagus on an average with a lower number of teeth. Very similar to C. hecla .....C, nastes Bdv. (PI. III:35"38). 12. Bigger, superuncus broader ......C. eurytheme Bdv. (PL 111:39, 40). C. philodice Gdt. (PI. III:4l). 12’. Smaller, superuncus slenderer. ....C. meadi Edw. (PL 111:42, 43). 13. Superuncus bent downwards C. phicomone Esp. (PL 111:44). 14. Superuncus diverging, very broad near the distal end. Clasper (in lateral view) broadest ventrally C. croceus Fourcr. (PL 111:45). 14’. Superuncus often with parallel or converging edges or clasper broadest in the middle part 15. 15. Clasperhead more strongly marked, tooth strongly developed. Ven- tral arm of aedeagus broad in the distal part. C. wiscotti Stgr. (PL 111:46, 47) C. marcopolo Gr. Grsh. probably also belongs here. No material has been available for investigation. Professor Sheljuzhko informs I (2); 1)5-156, 196) COLIAS GENITALIA 147 2/10 I CO 1 4/30 2/15 0 0 0 fO 0 CD CM 0 fO 0 CO CD to 0 0 s CO CO o .1^ o CO o CM o zS- - e. CM f—h - - c - CM - CO CM - eH iH - - - - CM CO - CO - - OJ CO CM rH t—i CM - - - - j C. heos 1 C. romanovi C. staudingeri C. thisoa C. wiscotti 1 C, vlluensis | j C, alpherakyi j C. christophi j C. cocandica i B 0 0 U j C, sieversi | C. sifanica Pop. of Table 3. C. aurorina C, sagartia j C« australis | C. hyale C. erate j C* rayrmidone j C. cunninghami ! C. vautieri | j P. imperialis j that Avinov and Kusnezov both place marcopolo together with wiscotti, the latter as a subspecies. According to Sheljuzhko both forms fly together in the Pamir area and are probably different species in spite of a great similarity in the genitalia. 15’. Clasper head less strongly marked 16. 16. A number of species remain which due to lack of sufficient material and/or great similarity could not be separated from each other : Yellow species with nastes-p2XtQtn\ C. dpherakyi Stgr. (PI. 111:48), C. chfistophi Gr.-Grsh. (PL IV:49), C, cocandica Erscii. (PI. IV: 50) (Specimen figured with triangular saccus; in a second specimen investigated the saccus is rounded as in all other species), C, sieversi Gr.-Grsh. (PI. IV:51), C. sifanica Gr. Grsh. (Pi. IV: 52). These Central Asiatic species together with C. nastes, phicomone, montium and others form one or — most probably — two polytypic species. 148 PETERSEN /. Res. Lepid. n 2 3 4 5 6 7 0 9 10 11 12 M C. hecla, N. Am. 8 5 1 2 6.6 *' , Greenl. 3 1 1 1 5.0 " , Scand. 30 2 7 9 8 4 6.2 C. meadi 10 4 1 3 1 1 7.6 C. eurytheme 12 3 3 5 1 6.4 C. palaeno, N. Am. 10 1 1 5 2 1 5.2 '* , Asia 8 1 1 6 5.5 " , Estonia 8 1 3 2 1 1 6.1 *' , Scand. 21 1 2 7 2 3 1 3 1 1 5.4 C. interior 11 3 4 2 2 5.3 C, Christina 1 1 6 C, nastes, N. Am. 6 1 4 1 6.2 *' , Asia 5 1 3 1 6.0 ” , Scand. 23 1 4 8 5 5 5.4 C. philodice 5 1 2 1 1 7.0 Table 5, Number of teeth near the apex of the aedeagus of Colias species with a slender superuncus. Orange species with hecla-pSLttern: C, chrysotheme Esp. (PI. IV:53), C, eogene Fldr, (PI. IV; 54, 55), C. staudingeri Alph. (PI. IV: 56), C. thisoa Men. (PI. IV:57), C. viluiensis Men., and C. heos (PI. IV:58, 59). As in the previous group it is un- certain whether all these forms deserve each a specific status. Orange species with «^j/^j-pattern : C. romanovi Gr. Grsh. (PL IV:60). This enumeration of Colias species is not complete. DISCUSSION The key given above does not show the phylogenetic relationships. These are, due to extensive introgressive hybridization, very difficult to find out. Suppose species A is closest related to the allopatric species B which has given rise to species C sympatric to A. Due to the introgressions A C or C A it is then possible that A and C are more similar in several characters than A-B or B-C. However, some characteristics, as for instance those on which the isolation between A and C depend, are likely to be of maximum difference in the case A-C. Among the Co/m-species investigated there exists a number of pairs of species within which the male genitalia are very similar: C. eurytheme-philodice; C. hecla-nastes; C. aufofina-sagartia; C. myrmi- done-erate; C. hyale-australis and probably also C. pdaeno-christina and C. wiscotti-mar copolo. Introgression has among these been established between C. eurytheme-philodice (Hovanitz 1949 a, b) and between C. hecla and nastes. Hybridization has been suggested between C. sugar- tia and aurofina (Lederer 1941). It is possible that pairs of a similar kind are present in Central Asia among the species figured from number (2);i}5-i56, 1963 COLIAS GENITALIA 149 48 to number 58. Lederer (1941) mentions, among other suspected hybrids, specimens which have been supposed to be hybrids between C. eogene and cocandica. A number of species with different genitalia have been seen in copula ( Lederer l.c. ) : hyde x myrmidone, hyde x croceus, hyde x erate, croceus x erate, and hyde x phicomone. The last cross gave rise to some larvae which died before the pupation. All these species have markedly different genitalia and are probably not very closely related. It therefore seems uncertain whether hybridization in these cases can give rise to introgression. If introgression occurs it must be possible to observe this fact also on the . genitalia. In all the pairs of sibling species first mentioned {eurytheme- philodice etc.), except hyde-austrdis and pdaeno-chfistina, colors, and in most cases also patterns, are different. The first species is usually orange with a heda-p2XXQm, the second yellow with a nastes-^2XX^iVi. Only C. sagarda has in addition a blue pigment which is present in some individuals of its orange sibling aurorina. The common occurrence of differences in color and pattern between sibling species suggests that these colors are integrating part of the isolating mechanisms within the pairs. Sexual isolation of this kind has in the pair Pieris napPbryoniae been established by Petersen, Tornblom and Bodin (1952). Males of both species are attracted by the white color of the P. napi female. The yellow female of bryoniae attracts the males solely by means of movements and odors. No releasing effect of any of the types of pattern was obtained neither in these pierids nor in similar experiments with the Silverwashed Fritillary {A. paphia L.) (Magnus 1954). Different color but probably not different pattern may therefore play a role for the sexual isolation between Colias species. The geographical distribution of some subdivisions of the genus Colias may also be discussed. Protocolias and Mesocolias are entirely South American. The species of the subgenus Colias with a broad superuncus are all Palearctic, one group having penetrated even into the Ethiopian region. Colias with a slender superuncus are mainly distributed in North America and the northern Palearctic, only a few living further south in the latter region. The distribution of the groups is to a great extent certainly the .. result of an evolution within different areas. It does not seem established that palearctic Colias have evolved from South American forms. They may as well have developed from primitive Colias in some other part of the world where they are now extinct. The evolution of the superuncus. Among the two characteristics of the Colias genitalia, the superuncus and the pseudouncus, the latter is present in the genus Anteos as well as in Catopsilia and Colias (Klots 1929 a, b, Drosihn 1933). The superuncus, on the other hand, is very small and triangular in the Anteos species investigated {menippe, clofinde, Plate IV, figs. 61, 62) covering only the pseudouncus and the basal part of the uncus. A similar, though bigger superuncus is found in Protocolias imperialis and the genus Phoebis (Drosihn 1933). In 150 PETERSEN /. Res. Lepid. the latter genus no pseudouncus is present. The Catopsilia and Aphrissa species have broad, straight and diverging superunci (cf. Plate IV, figs. 63, 64 and Drosihn 1933), rather similar to those of some Colias species. The triangular shape of the superuncus is probably primitive, as this shape is present in Anteos where the superuncus is comparatively small. The superuncus has in the genus Colias (and probably also in Catopsilia and Aphrissa) taken over the function of the uncus of many other Lepidoptera (cf. Lorkovic 1947) to assist medially and dorsally in holding the female body during the copulation. The superuncus and the pseudouncus probably developed to support the uncus dorsally. As the superuncus became larger, it was placed directly against the female body, and thus the uncus instead changed to support the super- uncus. It has been suggested, for instance by Verity (1947), that Colias have developed from the Old World Catopsilias. The presence of the very primitive Protocolias with a triangular superuncus, a short pseudo- uncus, and the middle part of the clasper with a prong, makes this assumption rather unlikely. The triangular superuncus may be con- sidered as a very pronouncedly convergent one ( cf . fig. 2 ) . The more or less parallel or divergent superunci of Colias, Catopsilia and Aphrissa may have evolved by means of parallel evolution towards more uncus- like conditions as discussed below in the case of Colias. The Catopsilias and most Colias species have broad, straight, and unpointed superunci, characteristics which therefore may be considered primitive compared with those of C. hecla, nastes and palaeno. A similiar result is arrived at, if the problem is approached from another direction. As already mentioned the superuncus has in the genus Colias taken over the function of the uncus. The latter has had its function during so much longer a time that it may give a certain indication of what shape is most apt to give an optimal function. The uncus is slender, pointed, and rather straight. In the two first of these character- istics the uncus corresponds to the superuncus of the hecla-nastes- palaeno-x^^t. Only it is even more pronounced then. The shape of the uncus on the other hand, is straighter in lateral view. The basal part of the superuncus is, however, situated more dorsally than the same part of the uncus ( cf . textfig. 13). Only if the superuncus is bent down- wards, it can be placed against the female body in the same place as the uncus of other Pierids. The evolution of the superuncus thus seems to have converged with that of the uncus. There is, however, still a marked difference in shape between these two organs and for different reasons it is not likely that the convergence will become complete ever. As already mentioned the situation of the two organs is different. The pressure of the super- uncus is supported by that of the uncus and probably also by that of the pseudouncus, and finally the claspers are in the genus Colias not built as in other genera. (2):i)5-i56, 1963 COLIAS GENITALIA 151 SUMMARY The male genitalia of a number of Colias species have been de- scribed. The genus is divided into one new genus and two subgenera of the genus Colias: The South American genus Protocolias (type imperialis Btlr.) and subgenus Mesocolias (type vautieri Guer.) and the mainly Holarctic subgenus Colias. The latter can be divided into a palearctic group with a broad straight superuncus and a nearctic and northern palearctic group with a slender superujicus which is bent downwards. Transitions between these two groups exist within the supraspedes C. nastes. The evolution of the superuncus from a small beginning, as at present in the genus Anteos, via the broad superund of the Catopsilias and some Colias into a slender superuncus has been discussed. Several pairs of Colias species with identical or very similar geni- talia exist. It is suggested that the similarity is combined with intro- gression. Several of the pairs include species of different colors. These colors may serve to isolate the species sexually from each other. REFERENCES DROSHIN, J- 1933. Uber Art-und Rassenunterschiede der mannlichen Kopulationsapparate von Pieriden, 1-135.. Stuttgart. HOVANITZ, W. 1949a. Increased variability in populations following natural hybridization. In Jepseo, et al. Genetics, Paleontology, and Evolution. 339-355. Princeton. HOVANITZ, W. 1949b. Interspecific matings between Colias eurytheme and Colias philodice in wild populations. Evolution. 3: 170-173. KLOTS, A. B. 1929a. The genus Anteos Hubner ( Lepidoptera, Pieridae). Bull. Brooklyn Ent. Soc. 24: 134-214. ^ 1929b. The generic status of Catopsilia Hubner and Phoebis Hubner, with a discussion of the relationships of the species and the homologies of the male genitalia (Lepidoptera, Pieridae). Bull. Brooklyn Ent. Soc. 24: 203-214. KUSNEZOV, N. J. 1915. Insectes Lepidopteres I in Faune de la Russie 1-337. Petrograd. LEDERER, G. 1941. Die Naturgeschichte der Tagfalter. Stuttgart. LORKOVIC, Z. 1928. Analyse des Speziesbegriffes und der Variabilitat der Spezies auf Grund von Uotersuchungen einiger Lepidopteren. Glasn. Soc. Scient. Nat. Croat. 38: 1-64, LORKOVIC, Z, 1947. Modes artifidels d’accouplement des papillons. Glasn. Biol. Sekc. Ser. II/B 1 :86-98. MAGNUS, D. 1954. Experimentelle Untersuchungen am Kaisermantel zur Analyse optischer Auslosungsreize. Deutscher Entomologentag in Ham- burg 30. Juli bis 3. August 1953: 58-75. PETERSEN, B., O. TORNBLOM, and N. -O. BODIN. 1952. Verhaltens- studien am Rapsweissling und Bergweissling (Pieris napi L. und Pieris bryoniae Ochs.) Behaviour. 4: 67-84. WARREN, B. C. S, 1950. Speciation in the genus Colias: with special reference to C. hyale and C. australis. Lambillionea. 50: 90-98. 152 PETERSEN /. Res. Lepid. PLATES a. lateral view b. superuncus, dorsal view c. caudal view d. aedeagus, ventral or dorsal view e. ” , distal end in lateral view f. clasper from inner side g. saccus in ventral view u. uncus v.l. ventral lobe of the inner side of the clasper (Klots 1929 a) PLATE I. Fig. 1. P. imperialis (wrongly labelled Honolulu) Fig. 2. C. vautieri, Ensenada Fig. 3. C. vautieri, Ensenada Fig. 4. C. ctinninghami, Junin, Peru Fig. 5. C. hyale, Oland, Sweden Fig. 6. C. myrmidone , Germany Fig. 7. C. sagartia, N.E. Persia Fig. 8. C. aiirorina, Armenia Fig. 9. C. montium, Tatsienlou, Tibet F.g. 10. C. moiitium, S. Kansu, China Fig. 11. C. interior, Montreal, Canada Fig. 12, 13. C. interior, Alaska Highway, mil 126, Beatton R. area, B.C. PLATE II. Fig. 14. C. interior, Alaska Highway, mil 90, Beatton R. area, B.C. Fig. 15. C. interior, Alaska Highway, mil 90, Beatton R. area, B.C. Fig. 16. C. palaeno, Alaska Highway, mil 450, Toal River, B.C. Fig. 17. C. palaeno, Bjurfors, Sweden Fig. 18. C. palaeno, Sweden Fig. 19, 20. C. palaeno, Smaland, Sweden Fig. 21. C. palaeno, Abisko, Sweden Fig. 22, 23. C. palaeno, 10 miles South Burwash Landing, Y.T. Fig. 24, 25. C. palaeno, Sweden Fig. 26. C. Christina, America borealis Fig. 27. C. Christina, Rocky Mts. Fig. 28. C. Christina, Am. bor. Fig. 29. C. hecla. Nr. Haines Junction, Y.T., up Summit Cr. 6000’-7000’ el. Fig. 30. C. hecla. Bog. nr. Johnson’s Crossing, Y. T. Fig. 31. C. hecla, Dudinska, Siberia Fig. 32. C. hecla, Sweden Fig. 33. C. hecla, Kvikkjokk, Sweden Fig. 34. C. hecla, 20 miles South Burwash Landing, Y. T. PLATE III. Fig. 35. C. nastes. Nr Haines Junction, Y. T., up Summit Creek, 6000’-7000’ el. Fig. 36. C. nastes, Mt. Atabaska, Jasper N.P., Alberta, 7000’-8000’ el. Fig. 37, 38. C. nastes, Kvikkjokk, Swedish Lapland Fig. 39. C. eurythemc, Texas Fig. 40. C. eury theme , Minnesota, U.S.A. Fig. 41. C. philodice, Amer. bor. Fig. 42, 43. C. meadi. Bow Pass, Jasper N. P., Alberta Fig. 44. C. phicomone, Alps Fig. 45. C. croceus, Tirol, Austria Fig. 46. C. wiscotti, Turkestan Fig. 47. C. wiscotti separata, Turkestan Fig. 48. C. alpherakyi , Turkestan PLATE IV. Fig. 49. C. christophi, Turkestan Fig. 50. C. cocatidica, Turkestan Fig. 51. C. sieversi, Turkestan Fig. 52. C. sifanica Fig. 53. C. chrysothemc Fig. 54. C. eogene, Kisil Fast area Fig. 55. C. eogene Fig. 56. C. staudingeri, Tian Shan, Fu-Shu-Shan Fig. 57. C. thisoa Fig. 58. C. heos, N. Mongolia, long. 100°- 110°, lat. 45°-50° Fig. 59. C. heos vespera, S. Kansu, China Fig. 60. C. romanovi, Turkestan Fig. 61. Anteos menippe, Mattogrosso Fig. 62. A. clorinde, Valles, Mex. Fig. 63. Catopsilia florella, Syria Fig. 64. C. grandidieri, Madagascar i(2):t35-i56, 196} COLIAS GENITALIA 153 PLATE I PETERSEN PLATE II COLIAS GENITALIA PLATE III 156 PETERSEN /. Res. Lepsd. PLATE IV 6fb\ Journal of Research on the Lepidoptera 1(2) :157-162, 1963 1140 W'. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 1965 P THE EFFECT OF HYBRIDIZATION OF HOST-PLANT STRAINS ON GROWTH RATE AND MORTALITY OF PIERIS RAPAE' WILLIAM HOVANITZ and VINCENT C. S. CHANG California Arboretum Foundation, Inc., Arcadia, Los Angeles State College and University of California, Riverside It has been previously shown (Hovanitz and Chang, 1962) that larvae of strains of Rieris rapae differ in their mortality rate, growth rate and final size dependent upon the plant which is supplied them for food. Slightly greater viability, greater rate of growth and increased size are characteristic results of feeding the larvae on kale (Brassica oleracea var. acephala) as compared with mustard (Brassica nigra), even though both plants are utilized by the species in the wild. It is the purpose of this paper to report on the results of an experi- ment designed to compare these physiological factors in strains bred for several generations on each plant 'and then hybridized. THE STRAINS The strains of Rieris rapae used in these experiments were derived originally from two sources, both in the Los Angeles Basin of southern California. The first, here designated the kale strain, originated with wild females obtained in a cabbage field (Brassica oleracea var. capitata) in a truck crop growing area in western Orange County, near Hunting- ton Beach. The other, here designated as the mustard strain, originated in the fields of the Los Angeles State and County Arboretum, where in the spring time Rieris rapae may be found in conjunction with black mustard (Brassica nigra). These have probably had no recent contact with cabbage since there are none grown commercially within nine miles of the area. Before testing in the experiments here described, the kale strain had passed through more than ten generations in the labora- tory on kale and the mustard strain had passed through more than six generations in the laboratory on mustard. 'Aided by a grant from the National Science Foundation, Washington, D. C. 157 158 HOVANITZ AND CHANG /. Res. Lepid. THE COMPARATIVE TESTS Six comparative tests were run contemporaneously, consisting of twenty larvae each. These larvae were fed in petri dishes by daily changes of food, keeping each larvae well isolated from the others. It was especially important while they were small that the larvae were placed directly on top of the leaf of the plant. Of the six tests, two were of the parent strains; one of these was derived from kale and fed on kale during the experiment, while the other was derived from mustard and fed on mustard during the experiment. Four further sets were Fi larvae from these two parent strains derived as follows: One set consisted of Fi larvae obtained by crossing a kale-strain female with a mustard-strain male and then by feeding the ensuing larvae on the food-plant of the mother ( kale ) . A second set consisted of larvae obtained by crossing a mustard- strain female with a kale-strain male and then feeding the ensuing Fi larvae on the food plant of the mother ( mustard ) . A third set consisted of larvae obtained as in the manner of the first F 1 set above, but fed on the food plant of the father ( mustard ) . A fourth set consisted of larvae obtained in the manner of the second F 1 set above, but fed on the food plant of the father ( kale ) . MORTAUTY The mortality figures for the larvae were low in all the strains ( table 1 ) . The kale-strain parents had a mortality of ten percent, which compared well with the data previously obtained ( Hovanitz and Chang, 1962). The mustard-strain parents had a mortality of thirty percent which also compared well with the previously recorded data. The reciprocal Fi’s behaved differently. When the kale-strain con- tributed the female parent and the mustard-strain contributed the male parent, the mortality of the Fi larvae was twenty percent whether kale or mustard was used as food. On the other hand, when the mustard- strain contributed the female parent and the kale-strain contributed the male parent, the larvae had only a ten percent mortality when used on kale and zero percent mortality when bred on mustard. It is doubtful that these slight differences are significant for the N value of 20, though they may represent a real difference. It is possible that the hybrids show a greater viability as a result of their heterozygosity. Both parental strains have been considerably inbred during their maintenance in the laboratory. No reason can be given at this moment for the greater viability of the mustard female F/s (10 and 0 percent) as compared with the reciprocal kale female F/s (20 percent each). i(2):t5y-i62, 196} PIERIS HYBRIDIZATION 159 LENGTH OF LARVAL PERIOD The minimum length of life for the larvae from the egg to the adult for both of the parental strains was twenty-five days under the conditions of the experiment (table 1) . All four of the hybrid F, strains had a shorter larval length of life indicating a greater activity than the inbred parental strains. The rate was speeded up to twenty-two days for the Fi strains in which the female parent was derived from kale stock. Type Strain No. No. No. and % Minimum days from egg to pupae Kale-bred 5 20 2 10% 25 Parents Mustard-bred 6 20 6 30% 25 u Kale-bred 2 20 4 20% 22 Kale ^ X Mustard ^ Mustard-bred 1 20 4 20% 22 u Kale-bred 4 20 2 10% 21 Mustard ^ x Kale 6^ Mustard-bred 3 20 0 0 23 Table 1, The comparative mortality and length of larval growth period of various strains of Pieris rapae. For the Fi stock in which the female parent was derived from mustard, the larvae bred on kale were speeded up to twenty-one days and those bred on mustard to twenty-three days. These data definitely indicate a significantly greater activity of the hybrid strains than the parental. However, there is no indication of any maternal effect on the inheritance of food-plant preference since the reciprocal crosses indicated the same general effects on mortality and growth rate whichever strain con- tributed the female parent. GROWTH RATE AND MAXIMUM SIZE The growth rate of the six sets of larvae was compared by measur- ing ten of the living larvae daily from the time of hatching from the egg to the time of pupation. The average of these ten measurements was plotted daily and curves drawn to illustrate their size increases (fig. 1). Differentiation between the curves occurred within the first few days but the first five days have been omitted in fig. 1, owing to the small size of the larvae. On the fifth day, both parental strains had the smallest larvae while the two F, strains, in which the kale-strain female parent was used, had the largest larvae. The other two Fi strains, in which the mustard- strain female parent was used, were intermediate. During the entire growth period, nearly the same relationships held^ with the exception of the kale-bred Fi with a mustard-strain female 160 HOVANITZ AND CHANG /. Res. Lepid. Fig. 1. Growth rates of larvae of six strains of Pieris rapae on kale and mustard. Vertical scale, length in mm.; horizontal scale, days after hatching from egg to pupation. (2):i5j~i62f 1^63 PIERIS HYBRIDIZATION 161 parent which for a time was faster than aM the others. Of the two parental groups, the mustard -bred larvae were larger than the kale-bred at five days, but the kale-bred reached a maximum ahead of and larger than the mustard-bred. Of the three groups on kale as compared with the three on mustard, the kale-bred group in each case reached a maxi- mum size larger than or earlier than the comparable group on mustard. Of the parental as compared with the Fi groups, the larvae of both F] hybrid groups consistently exceeded the larvae of the parental groups from the beginning to the end. The largest maximum size was attained by the F\ group in which the female parent was derived from a kale-strain and in which the larvae were kale-bred. On the other hand, the second largest maximum size was attained by the larvae of the F] group bred on mustard but which had a mustard-strain mother. It may be observed from the curves that at the periods of molting, sizes of the larvae are reduced. These periods are shown by the arrows (fig. 1). Because these reduction periods are delayed for the slower larvae, the older the larvae, the more difficult it is to compare the six groups on any one day. SUMMARY AND CONCLUSIONS 1. Crosses were made between strains bred for many generations on kale and mustard. 2. The effect of hybridization between these host-selection strains has been to speed up development, to bring about increased sizes to the larvae, and to reduce mortality. These effects are probably due to the genetic effects of "hybrid vigor”, in which the deleterious effects of inbreeding within the host-selected strains have been reduced by out- crossing. 3. These effects are greatest for the kale-strain than for the mustard- strain. This would tend to confirm our previous findings that- kale is preferred as a food over mustard, even when mustard is utilized as the food for many generations. 4. The increased viability, increased growth rate and increased size of the larvae are greatest when the hybrid larvae are grown on the plant corresponding to the strain of the mother, rather than the father. For example, the F-i hybrids where the mother was of the kale-strain grow larger and pupate sooner on kale than on mustard. Likewise, the Fi hybrids where the mother was of the mustard-strain grow larger but not faster on mustard than on kale. LITERATURE CITED HOVANITZ, W. aod VINCENT C. S. CHANG. 1962. The effect of vari- ous food plants on survival and growth rates of Fieru J. Res. Lepid., 1(1): 21-42. 162 HOVANITZ AND CHANG /. Res. LepiJ. NOTICE THE JOURNAL OF RESEARCH ON THE LEPIDOPTERA started publication in August, 1962. This fact makes it impossible for volume 1 to coincide with the calendar year, 1962 and still have four issues to the volume. The editor was confronted with tbe problem of two alternatives, ( 1 ) to extend the period for volume 1 to include the year 1963 and thus to have only four issues for the year 1962 and 1963, or (2) to have volume 1 completed early in 1963 by rapid issuance of the remaining numbers and then to issue the four numbers of volume 2 before the end of the year 1963. The second alternative has been selected, and therefore two more issues of the JOURNAL volume 1 will appear shortly, to be followed by four issues of volume 2 before the end of the year. Orders received for the calendar year 1963 with no volume designa- tion will be filled with volume 2 only; if volume 1 is desired, the year 1962 (volume 1) must be ordered. It is hoped that this explanation will help to solve the problems of institutional orders which have appeared. Orders for individuals who subscribe before the fourth number of volume 1 appears, will have their subscriptions start with the first volume, and will be billed for volume 2 at the time of sending number 4 of volume 1. In view of this rapid publication of papers, it is very likely that the editor will run short of manuscripts. It is urged that those who have good work to publish on the Lepidoptera will take advantage of this "chance of the lifetime” and send their MS to the editor. It is suggested that subscribers and readers help the JOURNAL by becoming a member of THE LEPIDOPTERA FOUNDATION. See inside front cover. Journal of Research on the ^epidoptera 1(2) ; 163- 168, 1963 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright CHANGE OF FOOD PLANT PREFERENCE BY LARVAE OF PIERIS RAPAE CONTROLLED BY STRAIN SELECTION, AND THE INHERITANCE OF THIS TRAIT’ WILLIAM HOVANITZ and VINCENT C. S. CHANG California Arboretum Foundation, Inc. Los Angeles State College, Los Angeles and University of California, Riverside It has been shown that larvae which have previously been fed on a particular food plant are more likely to select that kind of plant than if they had not previously been fed on it (Hovanitz and Chang, 1962 a and b). These data have shown that there was a shift in selection of mustard or of kale according to whether or not the larvae were from a kale or a mustard strain. Likewise, there was shown a greater selection for nasturtium by larvae previously fed on nasturtium, and indeed, by these same larvae, a greater selection for mustard over kale. This change in selection operates whether the larvae were fed on the changed food- plant for a long time (several generations) or a short time (part of the individual larval life) . In this paper, it is our purpose to test the effect of continued feeding of the larvae of Pieris rapae for many generations on mustard and kale by the process of isolating the strains so fed for many genera- tions, then crossing the parental strains to get Fi individuals and finally crossing the Fi to get the Fq segregation. THE STRAINS The strains of Pieris rapae used in these experiments were derived originally from two sources, both in the Los Angeles Basin of southern California. The first, here designated the kale strain, originated with wild females obtained in a cabbage field (Brassica oleracea var. capitata) in a truck crop growing area in western Orange County, near Hunting- ton Beach. The other, here designated as the mustard strain, originated in the fields of the Los Angeles State and County Arboretum, where in the spring time Pieris rapae may be found in conjunction with black mustard (Brassica nigra). These have probably had no recent contact with cabbage since there are none grown commercially within nine miles of the area. Before testing in the experiments here described, the kale strain had passed through more than ten generations in the labora- 1 Aided by a grant from the National Science Foundation, Washington, D. C. 163 164 HOVANITZ AND CHANG /. Res. Lepid, tory on kale and the mustard strain had passed through more than six generations in the laboratory on mustard. Larvae were selected from each and tested for their seleaion for a series of plants in the manner indicated in a previous paper (Hovanitz and Chang, 1962). The plants used in these tests were mustard (Brassica nigra), kale (Brassica oleracea var. acephala), nasmrtium (Tropaeolum majus), Isomeris (Isomeris arborea) and Cleome (Cleome huea). STRAIN SELECTIONS The tests were physically carried out in the manner indicated above. In order to be precise on the nature of any difference to be detected by these experiments, the tests were conducted in large numbers. Twenty- five larvae of each strain were used for twenty or forty trials each, giving a total of six hundred test times for each strain (table I). This was increased to six hundred sixty for the Fi and nine hundred (with 45 individuals) for the F2. mustard kale nasturtium Isomeris Cleome none no. of total larvae test PARENTS On mustard 364 60.66% 119 19.83% 56 9.33% 18 3.0% 12 2.0% 31 5.16% 25 600 On kale 144 24.00% 354 59.0 % 50 8.33% 21 3.5% 15 2.5% 16 2.66% 25 600 T M 9x K cf 356 59.93% 141 21.36% 90 13.63% 23 3.48% 39 5.91% 11 1.67% 25 660 (on mustard) 9 d- (on kale) 328 49.68% 127 19.24% 109 16.51% 33 5.00% 36 5.45% 27 4.09% 25 660 ^2 m'9x K d- 491 54.55% 145 16.11% 190 21.4 % 30 3.33% 42 4.66% 2 0.22% 45 900 (on mustard) K9 X M cf (on kale) 406 45.11% 146 16.22% 219 24.33% 79 8.77% 39 4.33% 11 1.22% 45 900 Table 1. The comparative mortality and length of larval growth period of various strains of Pieris rapae. The differential selection of the plants by larvae of the two selected strains is clearly made apparent by the curve showing the percentage selection (fig. 1). The larvae from the mustard bred strain selected mustard sixty-one percent of the time as compared with kale twenty percent of the time. The kale strain larvae selected kale fifty-nine percent of the time as compared with mustard twenty-four percent of the time. The two strains show an almost complete reversal of their i(2): i6j-i6S, 1963 PIERIS STRAIN SELECTION 165 preferences with regard to these two plants, their preference being in the direction of the plant utilized as food for several generations. The selection by larvae of these two strains for the other plants concerned was not significantly different. Nasturtium was selected nine percent of the time by the mustard strain and eight percent of the time by the kale strain. Isomeris was selected three percent of the time by each strain, and Cleome about two percent of the time. About five Fig. 1. Curves showing the percentage selection of various food plants by larvae of Pieris rapae from a kale and a mustard strain. percent of the larvae of the mustard strain left test area with no selection as did about three percent of those of the kale strain. Thus, there is no difference observable between these two parental strains in their selection, except with regard to mustard and kale. THE CROSSES Adults of the two strains were crossed and the larvae tested for their preference toward a selection of plants. Since there was a possi- bility of a maternal influence on the inheritance of the food plant selectivity hinted at in a previous experiment, reciprocal crosses were 166 HOVANITZ AND CHANG /. Res. Lepid. Fig. 2. Curves of the percentage selection of various food plants by larvae of Pieris rapae comparing the reciprocal F, of the strains shown in figure 1. made, mustard strain female X kale strain male and vice versa. In each case, however, the larvae were bred on the plant of the female strain until tested. It is now known that it would also have been desirable to have had the reciprocal feeding tests made in addition, as can be seen below. The F, larvae obtained, in which the female parent was from the . mustard strain, showed little difference in selection of plants from the parent mustard strain except in an increase in selection of nasturtium and Cleome, neither plant of which was involved in this selection. The reciprocal Fi larvae obtained from the cross of a kale strain female with a mustard strain male showed a much greater selection of mustard than the kale parent strain, and was more like that of either the mustard parent or of the reciprocal strain indicated (table 1, and fig. 2). In fact, the results of this cross would indicate that the genes for mustard selection are nearly completely dominant over those for kale selection. The cross would also indicate that they are transmitted at least through the male since a mustard strain male was used with a kale strain female and the larvae were bred on kale. The fact that the larvae were bred on kale may be the only reason that the selection was not higher toward mustard than is indicated. As with the reciprocal cross, there was a great increase in selection of nasturtium and Cleome following crossing. This was followed slightly by an increase toward Isomeris though this may not be significant. There is little doubt from these data that the selection of food plants is inherited, and that i(2):i6}-i6S, 196} PIERIS STRAIN SELECTION 167 "mustard” is dominant over "kale” despite the fact that there is also superimposed upon this inheritance a selective propensity controlled by training during the life of the individual, as has been shown before. The F2 crosses were made in the same manner. F2 larvae of each of the crosses indicated above were obtained and bred on the food plant of the mother ( table 1 ) . The tests on these larvae led to basically the same results as on the Fi larvae with the exception that all percentages were reduced slightly and the percentages for nasturtium were raised greatly. Of significance too is the reduction of the larvae which made no selection at all. The reduction in selection of mustard and kale appears proportional and related to the increase in the selection of nasturtium (fig. 3 ) . DISCUSSION AND CONCLUSIONS The data indicated in this paper appear to show that selection of genes for food plant preference occurs over a period of generations when strains are maintained isolated on particular food plants for many generations. These genes do not show maternal inheritance. In the crosses indicated in this paper, genes for selection of mustard over kale are apparently dominant over the reverse. This is indicated by the fact that in the Fi of the cross kale strain X mustard strain or the reciprocal, selection is similar to that shown by the original mustard strain. The same results are shown for the F^ but they are not so pronounced. Of great curiosity is the fact that in the Fi there is an increase in selection for nasturtium, not one of the preferred plants of Pieris rapae ( Flovanitz and Chang, 1962a). Feeding larvae on nasturtium has shown that there may be greater selection of this plant developed than would ordinarily be present. In the present case, however, no selection of this sort is involved and the increase is significant. This increase is even greater in the F2 cross than in the Fi, there being a two and one half- to three- fold increase in selection of nasturtium in the F2 as compared with the parental strains. A much slighter increase is also indicated for Isomeris and Cleome, and a decrease for the number of rejects, that is, those that make no selection. Reasons for this much greater selection for nasturtium and other plants following hybridization are not known nor can they even be guessed at intelligently at this time. SUMMARY 1. Two food plant strains were developed by selection, one on kale (Bras sic a oleracea) and another on mustard (Brassica nigra). 2. When tested, each of these strains showed a much greater preference for their accustomed food plant than for any other tested. Those tested were mustard, kale, nasturtium, Isomeris and Cleome. 3. When these two strains are crossed, the Fi hybrids showed a preferential selection most like the mustard parent strain. This indicates that the gene(s) for mustard is dominant over that for kale. There is 168 HOVANITZ AND CHANG J. Res. Lepid. Fig. 3. Curves of the percentage selection of various food plants by larvae of Pieris rapae comparing the reciprocal F2 of the strains shown in figures land 2. also indicated a slight increase in selection of the other plants tested. 4. The Fo of this cross showed results similar to the Fi, namely a selection in favor of mustard rather than kale whether or not the female parent had originally come from the mustard or the kale strain. Thus, maternal inheritance is not indicated here even though maternal effects were indicated on growth rate and mortality data in a previous paper. 5. As had been indicated, feeding the larvae even a short time on one particular food plant may influence it to have a preferential selection for that plant. 6. In the F2 there is indicated a strongly increased preferential selection toward nasturtium which was not selected in any previous strain. This increase is also indicated slightly toward Isomeris and Cleome. No reason is advanced at this time for these results. LITERATURE CITED HOVANITZ, W. and VINCENT C. S. CHANG. 1962a. The effect of vari- ous food plants on survival and growth rates of Pieris. J. Res. Lepid. 1(1): 21-42. 1962b. The effect of hybridization of host-plant strains on growth rate and mortality of Pieris rapae. J. Res. Lepid. 1(2) : this issue. advertisement TAKi CHLOROCRESOL ON YOUR NEXT TRIP Since the publication of the Chlorocresol method, field collecting techniques have been changed radi- cally. The Chlorocresol method eliminates the need for time consuming field mounting and/or relaxing of specimens. With Chlorocresol, specimens are re- tained in a relaxed condition for an unlimited period, allowing mounting to be done at a more convenient time or place. Field mounted specimens which would normally occupy a standard insect box may be packed in a container only 5" x 5" x IV2'. With specimens in a relaxed state, damage is minimized in the event of rough handling. Catalog No. 182 Chlorocresol, 50 grams $0.95 Chlorocresol, 100 grams 1.75 Chlorocresol, 200 grams 3 00 Styrene box, 5" x 5" x IV2' per dozen 3.00 Complete instructions are included with each shipment. Prices are F.O.B. Santa Monica, California. California residents please add 4% sales tax. General Entomological Supplies BIO METAL ASSOCIATES Box 61 Santa Monica, California Volume 1 Number 2 January, 1963 THE J@U^NJAL ©F RESEARCH ©NJ THE LEFIJ©©RTERA\ IN THIS ISSUE Composition and relative abundance in a temperate zone butterfly fauna . . . Thomas C. Emmel and John F. Emmel 97 The Argynnis populations of the Sand Creek area, Klamath Co., Oregon, Part I J- W. Tilden 109 Caterpillar Versus Dinosaur? . . . Theodore H. Eaton, Jr. 114 Geographical distribution and variation of the Genus Argynnis I. Introduction II. Argynnis idalia William Hovanitz 117 The relation of Pieris virginiensis Edw. to Pieris napi L. Species formation in Pieris? Wiliam Hovanitz 124 The male genitalia of some Co lias species . Bjorn. Petersen 135 The effect of hybridization of host -plant strains on growth rate and mortality of Pieris rapae William Hovanitz and Vincent C. S. Chang 157 Notice-editorial 162 Change of food plant preference by larvae of Pieris rapae controlled by strain selection, and the inheritance of this trait William Hovanitz and Vincent C. S. Chang 163 @F OIMJ THE LEFI1P@PTER1A MeSfeii Bi2!!*S*****»»*«'*" . ... ^^■rs****** ***•*•» !psr.i:j2:r*fe; us:sss:2s;*.2ui,. K«sjssK2.»:2i;ir Volume 1 Number 3 March, 1963 i, rr V * Established in 1 962 Published at 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. Edited by WILLIAM HOVANITZ WITH EMPHASIS ON ENVIRONMENTALLY AND GENE- TICALLY INDUCED VARIATION, population analysis, evolution, phylogenetic taxonomy, zoogeography, comparative morphology, ecol- ogy, geographical variation, speciation, physiology, etc. In short, quality luork on any aspect of research on the Lepidoptera. THE PURPOSE OF THE JOURNAL is to combine in one source the work in this field for the aid of students of this group of insects in a way not at present available. The JOURNAL will attempt to publish primarily only critical and complete papers of an analytical nature, though there will be a limited section devoted to shorter papers and notes. RATES: $ 8.00 per volume, personal subscription. $12.00 per volume, institutional subscription. All amounts are in U.S. dollars, payable in the U.S.A. AUTHORS ARE REQUESTED to refer to the Journal as an example of the form to be used in preparing their manuscripts. Fifty separates will be supplied to authors free; reprints will be sold at printer’s rates if ordered at time galley proofs are returned. If proofs are not returned promptly, the editor reserves the right to withhold publication, or to proceed with no responsibility. All issues of the Journal will be copyrighted; in submitting a paper each author agrees that the material has not been published elsewhere. MANUSCRIPTS AND SUBSCRIPTIONS should be mailed to the address noted above. THE LEPIDOPTERA FOUNDATION To provide permanent security to the JOURNAL, THE LEPIDOP- TERA FOUNDATION is being set up. Memberships are now being accepted at TEN DOLLARS per year, eight dollars of which goes to your subscription and two dollars goes into a permanent capital fund, the earnings only of which may be used for current publishing expenses. In addition a scale of higher contributions is to be set up which will be announced later. All funds received in excess of the subscription rate will go into the permanent fund. LIFE MEMBERSHIPS are offered now at Two hundred fifty dollars; these include the Journal for life. INSTITUTIONAL AND COMMERCIAL DONATIONS AND MEM- BERSHIPS are also available; rates will be announced later, or on request. These contributions are deductible. HELP YOUR JOURNAL. Meetings of members may be held at times and places to be arranged. Journal of Research on the Lepidoptera 1 ( 3 ) :169-182, 1963 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 196} SELECTION OF ALLYL ISOTHIOCYANATE BY LARVAE OF PIERIS RAPAE AND THE INHERITANCE OF THIS TRAIT' WILLLIAM HOVANITZ and VINCENT C. S. CHANG California Arboretum Foundation, Inc., Arcadia, Los Angeles State College and University of California, Riverside Larvae which have previously been fed on a particular food plant are more likely to select that plant than if they had not previously been fed on it (Hovanitz and Chang, \9G2a and ^). It was shown that larvae which were fed on kale are more likely to select kale, and larvae which were fed on mustard are more likely to select mustard. Also, larvae which have previously been fed nasturtium have a higher selection on that plant than if they had not been fed on it. This change in food plant selection operates both if the adaptation comes over a period of several generations, or if only for a short time in the individual larval life. Later, it was shown that strains selected for several generations on each of two plants, kale and mustard, had strong selection for their respective food plants ( Hovanitz and Chang, 1963^^ and b). Individuals from these two strains were then crossed to get the Fi generation, and these were crossed to get the Fo generation. The inheritance of this trait appears to be multi factorial but because of the nature of the testing, this point may be open to question. The data to be presented in this paper were obtained in much the same manner as in the previously mentioned papers (Hovanitz and Chang, 1963<? and b) with the exception that instead of testing the selection of the larvae for the plants themselves, they were tested for their selection of various concentrations of black mustard oil (allyl isothiocyaoate ) . Tests were made on the parental larvae as well as the Ft and the F2. THE STRAINS The strains of Pieris rapae used in these experiments were derived from two original sources. The first, the kale strain, originated with wild females obtained in a cabbage field {Brassica oleracea var. capitata) in a truck crop growing area near Huntington Beach in Orange County, California. The other, the mustard strain, originated in the fields of the 'Aided by a grant from the National Science Foundation. 169 SlllTflSIJNlW mSTlTUTiaH APR1819® 170 HOVANITZ AND CHANG /. Res, Lepid. Los Angeles State and County Arboretum where P. rapae is found on black mustard (Brassica nigra). Before testing in the experiments here described, the kale strain had passed through more than ten generations in the laboratory on kale {Brassica oleracea var. acephala) and the mustard strain had passed through more than six generations in the laboratory on black mustard. EXPERIMENTAL SET-UP The tests of the larvae were made in much the same manner as has previously been described (Hovanitz and Chang, 1962b) except that in place of the plants to be selected, at various points surrounding the flat, filter papers wetted with various solutions of allyl-isothiocyanate were placed. Preliminary tests had indicated that the larvae were attracted to dilutions of the mustard oil at 10"^ to 10"'^ so the tests were made at these dilutions. Metal mosquito netting was cut in a circular arrangement to serve as a base for the filter paper. These discs of metal netting were placed on top of glass flasks partially buried in soil, secured to corks placed in the flasks and arranged in various parts of the flat according to the disposition shown in Figure 1. The filter paper was then set up top of the metal netting and wetted with distilled water solutions as shown. On one of these, distilled water only was placed so as to serve as a control for the others. To prevent the accumulation of mustard oil concentrations, the solutions were alternated around the flat. The set-up of the experiments is shown diagrammatically in Figure 1 and photographically in Figure 2. The larvae were set in the center of the flat and allowed to select any of the six test papers or to leave the flat without selection. It should he noted that the test papers were above the level of the larvae and therefore that the heavier-than-air mustard oils would drop down to their level. After each trial, the soil in the flat was mixed to prevent the origin of "trails.” It should also be noted that it would be much easier for the larvae to leave the flat without any selection except for one factor, namely, that these larvae have a tendency to climb upwards when at all possible. Experience with lepidopterous larvae has shown that descent is usually made by means of dropping down on a silken thread, rather than by walking while ascent is made by walking. This accounts for the selection of the "water” alone just as frequently as they leave the flat without any selection at all. The mustard oil used in these experiments was a natural product obtained from black mustard seed by compression and distillation. It is available commercially from the Fritsche Brothers firm. Allyl iso- thiocyanate is the principal component of the oil obtained from Brassica nigra; other mustard oils are found in other plants and all are esters of isothyocyanic acid. Brassica nigra seeds contain a glucoside, potassium myronate or sinigrin, which undergoes fermentation change due to the enzyme myrosin present in the seeds. (})ti69-ii2, 196} ALLYL ISOTHYOCYANATE 171 Fig. L The set-up for the larval tests. The larvae are placed in the center of the flat from where they are enabled to go in any direction toward one of the dilutions of mustard oil, or to leave the flat. The circles show the disposition of the filter papers which have upon them dilutions of mustard oil as indicated. THE TEST SELECTIONS Test selections were made by using a number of larvae for several times to get a combined sample. The number of individuals used depended in part on those available, but ranged from fifteen to forty-five giving a total number of tests from three hundred to nine hundred. These data are given in Table 1. The parent strains were tested first. The mustard strain was tested with twenty -five larvae 20 or 40 times each giving a total of seven hundred trials. Of these, only nineteen or 2.7 percent made no selection and left the flat. Only twenty-four or 3.4 percent selected the filter paper with distilled water only on it. The others all selected the mustard oils, from a low of nearly 10 percent for the 10"’ to a high of nearly thirty percent for the 10"^. These data are shown diagrammatically in Figure 3. 172 HOVANITZ AND CHANG /. Kes. Lepid. Fig. 2: Photograph showing the surface of the experimental flat, the flasks with filter paper upon them and a larva raising its foreparts. The rules are not present during an experiment and are indicated only for the purpose of illustrating the dimensions between the soil surface, the filter paper and the larvae. for different concentrations of allyl isothiothiocyanate (black mustard oil). 1(3)^169-182, 1963 ALLYL ISOTHIOCYANATE 173 5 174 HOVANITZ AND CHANG /. Res. Lepid, The kale strain was also tested with twenty-five larvae for a total of five hundred eighty trials. The highest selection of these larvae was toward the concentration of 10"®, larvae going to this concentration in thirty-four percent of the trials. This result is significantly different from the result for the kale larvae indicated above, in which the highest selection was toward the 10"® concentration. These data can be com- pared diagrammatically in Figure 3. There is little doubt that the larvae of the kale-bred stock prefer the mustard oil at a lower concentration than do those of the mustard strain. It is probably not coincidental that the taste of the leaves of mustard is stronger than is that of kale (or cabbage) since the significant taste of these plants is due to allyl or other isothiocyanates. These data would seem to indicate ( 1 ) that the concentration of taste or smell-detectable mustard oils in the leaves of mustard is higher than that in kale, (2) that the larvae feeding on mustard have become adapted in some way to a higher concentration of these oils, and (3) that due to this adaptation, the "mustard” larvae have a greater selection of the higher concentration of mustard oil than the lower. It has been shown in another paper (Hovanitz and Chang, 1963b) that these two strains also select their own plants over the others when given a free choice. This preferred selection may be due wholly or in part to the difference in available allyl or other isothiocya- nate in their leaves. THE CROSSES The crosses between the two food plant strains were carried out by making reciprocal hybrids. In the one case, the kale strain was used as the female parent and in the other case, the mustard strain. The hybrid larvae resulting from these crosses were fed on the food plant of the mother, since the mother was allowed to lay on its own preferred plant. The selections made by the hybrid larvae in both crosses favored the concentration 10"^, exactly half way between the selections made by the respective parent strains ( Fig. 4 ) . The frequency of selections for the kale female Fi toward the 10"^ concentration was 34 percent as compared with 30 percent for the mustard female Fi. Previous data have indicated that the immediate larval food plant eaten by the larvae had an effect on the food plant selections made by these larvae. If this held true for the selections of the mustard oils also, then the larvae of the kale female Fi, having been fed on kale during their life, and the larvae of the mustard female Fi, having been fed on mustard during their life, should have a preference for their respec- tive plant. Data from Table 1 (Hovanitz and Chang, 1963b) indicate that this is so. Mustard female Fi larvae selected mustard nearly 60 percent of the total selections as compared with nearly 50 percent selec- tions for the kale female Fi larval selections. It should follow, if our PERCENT SELECTION PERCENT SELECTION i(3):i69-az, 196) ALLYL ISOTHIOCYANATE 175 Fig. 3: Histograms illustrating the selections made by the larvae from the parental strains, comparing the kale and the mustard strains. PERCENT SELECTION PERCENT SELECTION 176 HOVANITZ AND CHANG /. Kes. Lepid. Fig. 4: Histograms illustrating the selections made by the Fi from the reciprocal crosses between the kale and mustard strains. larvae PEHCEMT SELECTiON PERCENT SELECTION (3);i6^-ii2, 1^63 ALLYL ISOTHIOCYANATE 177 Fig. 5. Histograms illustrating the selections made by the Fa from the reciprocal crosses between the kale and the mustard strains. larvae 178 HOVANITZ AND CHANG /. Res. Lcpid. assumptions are correct concerning the relationship between mustard oil concentration and food plant selectivity, that the mustard female Fi larvae should select mustard oil concentrations at a higher level than the kale female F] larvae. That this is so can be seen by study of the histograms (Fig. 4) and the following calculations. Selections made by the mustard female Fi larvae which are less than 10"^ (that is 10"^ or ‘^) are 31.4 percent of the total as compared with 24.6 percent of the total for the kale female F, larvae. This shows that the larvae of the mustard female Fi select a higher concentration of mustard oil than those of the kale female Fi even though the modal class in both cases is at 10"^. This difference is most probably attributed to the fact that the immediate F, larvae had been fed on the food plant of the mother, rather than that there was any genetic effect of maternal inheritance. Two sets of Fq have been bred, corresponding to the two sets of Fi. These two sets of Fi are the reciprocal crosses mentioned above. The Fq from the mustard female Fi was continued on mustard for the food plant of the F2 larvae, and the F2 from the kale female Fi was con- tinued on kale. It would be expected, therefore, that each of these strains might diverge slightly in their selection according to the pattern of the larval foods. This result did not materialize. The data indicate that the Fq of both strains still have a selective propensity favoring the dilution, 10"^, thus having the same modal class as the F 1 ( Fig. 5 ) . This is the expected genetic result for multiple factor inheritance. The second result expected for multiple factor in- heritance is a wider range of variation in the F2 than the Fi. This result is not shown by a wider range per se since no wider range of dilutions were used, but rather it is shown by a greater dispersion of the selections inside the range tested. This would give a greater standard deviation for the F2 than for the corresponding Fi. The greater disper- sion is very clear by study of the histograms, rearranged so that the mustard strains are on one page and the kale strains are on another (Figs. 6 and 7). The differences between the Fi and the F2 are striking. These results indicate clearly that the factors for food plant selection in Pieris rapae which have been separated in the two strains here tested are genetic and behave in a way comparable with polygenic inheritanp. It also indicates that the populations of Pieris rapae existing in the wild are differentiated into genetically distinct populations, differing by genes which control the selection of food plants. These differences could account for a mechanism needed for species isolation (Hovanitz 1963c). DISCUSSION AND CONCLUSIONS The data indicated in this paper show that the strains which have been selected for particular food plant preferences over a period of several generations, and which when tested prefer the food plant upon which they have fed, also differ in their selection (preference) for vari- ALLYL ISOTHIOCYANATE 179 ous concentrations of mustard oil (allyl isothiocyanate). That these differences are genetic is apparent not only from the preceding data on food plant preferences but also on the data on mustard oil preferences introduced in this paper. Crossing of the two strains discussed here gives an F, which has a selection exactly intermediate between the two par- ental strains. The results are similar whether the kale strain contributes the maternal parent or whether the mustard strain contributes the maternal parent. There is, however, a greater selection of the higher concentrations in the strain fed on mustard than the one fed on kale. The F2 from these strains is similar to the Fi except for the wider dispersion of the selections. In the testing of these strains for food plant preferences, it was found that the larvae developed a preference for nasturtium in the Fi and F2. No reason can be deduced for this strange event based upon the present results unless it be that nasturtium emits at a concentration of mustard oil of the same attractive power as 10"^ allyl isothiocyanate or some other mustard oil of attractive significance. The odor of nasturtium does not seem to the human to be the same as that of allyl iosthiocyanate. However, it is possible that to the human, the odor of some other sub- stance masks the odor of the attractive mustard oil, but that this odor does not do so in the larvae which are able to detect it. Thorsteinsen (1953) has made the statement that larvae of the diamond-back moth, Plutella macuUpennis (Cut.) are attracted to the odor of ally! isothiocyanate from a distance but are repelled by it when closeup so that they will not eat substances having this oil. Instead, they will eat substances with the corresponding glucoside, potassium myronate, but are not attracted to this substance from a distance. Since the glucoside is converted into the allyl isothiocyanate in the plant by a process of fermentation, it is probable that the attraction is really for the isothiocyanate not the glucoside, but that the isothiocyanate is pro- duced slowly and in small quantities by the process of enzymatic fermentation. Indeed, Thorsteinson even indicates "in some experi- ments the addition of allyl mustard oils slightly increased feeding on media containing sinigrin.” Wolfrom (I960) points out that "while the exact biological function of the plant glycosides is not established, it is probable that their formation provides the plant with a means of storing, in a harmless form, toxic and physiologically active materials which may be liberated by enzymes, in small quantities, when required.” It seems not only possible but highly likely that the attraction differ- ences noted by Thorsteinson are based upon the low concentration of allyl isothiocyanate rather than upon the presence of the glucoside sinigrin directly. Thorsteinson in fact admits this possibility in his com- ment, "On the other hand, it is possible that infinitesimal amounts of mustard oil vapor emanating from the leaves [by enzymatic action of myrosin on sinigrin] may stimulate the olfactory sense which is char- acteristically extremely sensitive in insects.” Section in brackets not PERCENT SELECTION PERCENT SELECTION 180 HOVANITZ AND CHANG /. Res. Lepid. Fig 6. Histograms illustrating the selections made by the Fi and F:; larvae, comparing these larvae of the kale female crosses. PERCENT SELECTION PERCENT SELECTION (}):i69-i82, 196} ALLYL ISOTHIOCYANATE 181 Fig. 7. Histograms illustrating the selections made by the Fi and F2 larvae, comparing these larvae of the mustard female crosses. 182 HOVANITZ AND CHANG /. Res. Lepid. Thorsteinson’s. Further testing of feeding responses of Plutella or Pieris larvae at a low mustard oil level are required to be certain of this differentiation. Presently published data of the present, or of other, authors are definitely not conclusive. Tests of this sort may be very difficult to make due to the high volatility of the mustard oil. There is no doubt that the use of the glucoside, rather than the oil, makes easier the maintenance of a low level of the oil vapor in the vicinity of the food. Our own data comparing the egg laying of the adults which are to be published later shows that the glucoside and mustard oil released from crushing seeds is more attractive than the mustard oil alone. Verschaffelt (1910) was the first to study the relationship between the food plant choice of insects, plant odor and the chemistry of the relationship. He supposed that the larvae were attracted to the mustard oils present in the food plants but did not know whether the larvae differentiated between different oils or not. Dethier (1941) showed the selective effect of various odiferous oils toward lepidopterous larvae. Johansson (1951) showed that Pieris hrassicae larvae preferred food plants to which they had already become accustomed by previous eating. This result has also been shown by us (Hovanitz and Chang 1962b and 1963a) on Pieris rapae. LITERATURE CITED DETHIER, V. G. 1941. Chemical factors determining choice of food pianis by Papilio larvae. Amer. Nat. 75: 61-75. HOVANITZ, WILLIAM and VINCENT C. S. CHANG. 1962a. The effect of various food plants on survival and growth rate of Pieris. /, Res. Lepid. 1.(1): 21-42. — — ■. 1962b. Three factors affecting larval choice of food plant. J. Res. Lepid. 1(1): 51-62^ — — — ■. 1963a. The effect of hybridization of host-plant strains on growth rate and mortality of Pieris rapae. J. Res. Lepid. 1 (2) : 157-161. • — . 1963b. Change of food plant preference by larvae of Pieris rapae controlled by strain selection, and the inheritance of this trait. /. Res. Lepid. 1 (2) ; 163-168. . 1963c. The relation of Pieris virginiensis Edw. to Pieris napi L. Species formation in Pieris? /. Res. Lepid. 1 (2) ; 124-134. JOHANSSON, ARNE SEMB. 1951 The food plant preference of the larvae of Pieris hrassicae L. Norsk Ent. Tid. B. VIII, 2.4-5: 187-195. THORSTEINSON, A. J. 1953. The chemotactic responses that determine host specificity in an oligophagous insect {Plutella maculipennis (Curt.) Lepidoptera ) . Canad. J. Zool. 31: 52-72. VERSCHAFFELT, E. 1910, The cause determining the selection of food in some herbivorous insects. Proc. Acad. Sci. Amsterdam 13(1): 536-542. WOLFROM, M. L. I960. Natural glycosides. Encyclopaedia Britannica 10: 448.. Journal of Research on the Lepidoptera 1 (3 ); 183=190, 1963 1140 W. Orange Grove Ave., Arcadia, California, U.S.A. © Copyright 1^6} BIOLOGY OF THE CEANOTHUS STEM-GALL MOTH, PERIPLOCA CEANOTHIELLA (Cosens) With Consideration of its Control* J. ALEX MUNRO Los Angeles State and County Arboretum Arcadia, California The stem-gall moth, Periploca ceanothiella {Cosens), has received but slight attention considering its importance as a pest of ornamentals on the genus Ceanothus. The genus Ceanothus, according to Van Rensselaer and McMinn (1942), comprises some 55 species native only to North America. Of these, some 44 species, and a number of horticultural varieties or selections, are grown in the Pacific Coast states. Many of them exhibit rather colorful bloom, ranging from the white flowers of C. americanus (New Jersey Tea) to the deep blue, lilac -like clusters of C. thyrsifloms (Blue Blossom) and the pink of some selections. They occur mainly as shrubs and are used in land- scape plantings and roadside beautification. Some, such as C. gloriosus exaltatus (Point Reyes Creeper) and C. griseus horizontalis (Carmel Creeper) are used as attractive groundcovers because of their low growing habit. The distribution of the gall moth probably corresponds to the range of its host. Hodges (1962) in reviewing the genus Periploca mentions having examined specimens of P. ceanothiella collected in California, Kansas, New York, Ontario (Canada), and Texas. His review shifted this insect from the genus Stagmatophora, which it formerly occupied, to its present position in the genus Periploca. DEVELOPMENT This insect normally overwinters within its gall in some stage of larval development. Reference to Table 1 shows pupation to be barely under way in April and to continue, with moth emergence, throughout spring and early summer. Emergence, however, may begin as early as January from mature larvae brought in from outdoors and maintained under room temperatures. The small, dark-colored moths (Fig. 1) begin egg laying within a few days following emergence. They deposit their small (0.5 mm) white ■^Presented at the annual meeting of the Entomological Society of America, Phoenix, Arizona — December 3-6, 1962. 183 184 MUNRO /. Res. Lcpid. TABLE 1. Status of overwintered Periploca ceanothiella galls. PLACE 6 DATE PERCENT GALLS 1962 Number of galls exam With Larvae With Pupae With Parasites Emerged Arcadia April 9 88 87 2 11 0 May 18 64 60 21 15 4 June 7 142 23 44 19 14 July 26 104 2 6 3 89 Claremont April 25 85 52 28 18 2 May 16 80 28 22 11 39 June 14 76 19 26 15 40 La Canada April 17 66 94 0 6 0 May 22 70 62 4 19 15 Santa Barbara April 14 141 95 1 4 0 eggs on the underside of leaves, in leaf axils, and along the twigs. The eggs required 10 days for hatching under indoor temperatures of about 70° F. The newly hatched larvae move directly into the terminal growth where they penetrate the buds and inflorescences to cause the swellings or galls. Here the larva remains, feeding and growing on the inner tissue, until reaching full growth of 5 to 6.5 mm by the following spring or summer. Before pupation begins, the larva cuts an exit hole partially through the side of the gall to leave only a thin film of bark which covers the entrance until time for the moth to emerge. This "window” or blocked entrance provides protection for the occupant against some enemies until it is ready to escape from the gall cavity. The pupation stage is spent within the gall and requires 24 days under room temperatures of about 70 °F. Although the gall moths may emerge over a prolonged period of spring and summer, only one brood or annual generation was observed. The pest has a stunting effect on the new growth. The galls may vary greatly in abundance, ranging from only an occasional gall to 20 or more per lineal foot of branch and twigs in heavy infestations. Some (^)-.iS3-i90, 19^3 GALL MOTH 185 of the twigs may be killed outright from the injury. The most serious damage, however, was to the bloom. (Fig. 2). It was not unusual to see the flower clusters reduced to about 25% of their normal size, particularly where the gall larvae were present in the inflorescenses. In no instance was more than one larva found per gall. In appear- ance, the gall is somewhat spindle-shaped and averaged about 13mm in length by 6mm diameter. (Fig. 3) This is about three times the diameter of the twig at point of junction with the gall. PARASITES The role of three naturally occurring ichneumonids in checking the stem gall is indicated in Table 1. All three, Fristomems haumhoferi Cush., Scambus aplopappi (Ashm.), and Apistephialtes nucicola (Cush.), were observed mainly during the spring months. No hyper- parasites were observed. Identifications were made by L. M. Walkley of the U.S. National Museum Staff. Muesebeck, Krombein, and Townes (1951) mention these ichneumonids as parasites of certain micro- lepidopterous larvae that tunnel in stems. Of the three parasites, P. haumhoferi was the most commonly encountered, and was the only one previously recorded as a parasite of this stem gall insect. The toll taken by the parasites ranged up to 19% of the gall-larvae. Only the mature larvae were attacked by the parasites. NON-SUSCEPTIBILIl’Y IN CEANOTHUS Examination of Ceanothus collections at the Los Angeles State and County Arboretum in Arcadia, Rancho Santa Ana Botanic Garden at Claremont, and the Santa Barbara Botanic Garden at Santa Barbara (all southern California locations) showed varying degrees of non- susceptibility to the stem galls. It was an ideal way to check on this factor inasmuch as the Ceanothus species and varieties growing in each collection were equally exposed to the pest. The results presented in Table 2 would suggest the use of non-susceptible Ceanothus in plant- ings wherever practicable. CHEMICAL CONTROL Inasmuch as spraying of the Ceanothus with DDT had given little or no apparent control of this pest it was decided to test the systemic insecticide, dimethoate, for the purpose. A commercial formulation (Cygon) containing 43.5% dimethoate was mixed at the rate of one pint of this concentrate to 100 gallons of water and applied as a foliar spray to Ceanothus plantings in Arcadia on June 27, 1962. Examina- tion made after one week following application showed that 92% of 186 MUNRO /. Res. Li pid. TABLE 2, Occurrence of Periploca ceanothlella on various species, varieties and horticultural selections of Ceanothus. HEAVY OCCURRENCE; Ceanothus griseus C, griseus horizontalis MODERATE OCCURRENCE; C. cyaneus C, thyrsiflorus C, 'Ray Hartman* C, 'Marie Simon' LIGHT OCCURRENCE: C. arboreus C, oliganthus C, divers if olius C. lemmonii C, integerrimus C, leucodermis C, lobbianus C, 'Treasure Island* C. 'Sierra Blue* C, 'Royal Blue* C, 'Mountain Haze* C, 'Mary Lake' C. 'Concha* NO OCCURRENCES C, americanus C, parry i C. papillosus C, impressus C, foliosus C. insularis C. cuneatus C« ramulosus fascicularis C, gloriosus C. gloriosus exaltatus C. rigidus albus C. purpureus C. prostratus C. verrucosus C. spinosus C, jepsonii C, raasonii C« megacarpus C, greggi perplexus C, 'Blue Cloud* C. 'Lester Rowntree* the gall larvae were killed. The spray, however, had no appa'rent effect on the mature gall larvae (which had ceased feeding), the pupae, or the parasites, P. baumhoferi and A. nucicola; adults of these normally emerged from galls collected from the treated areas. Effect of the treat- ment on 5’. aplopappi was not observed. The parasites developed only in the mature gaU larvae, and therefore it is unlikely that they had had any contact with the insecticide. In addition to checking the stem gall infestation, the dimethoate also controlled the Ceanothus leaf miner, Nepticula ceanothi Braun, thrips, and an infestation of mites which were troublesome at the time. The spray used at this rate had no ob- servable phytotoxic effect on the Ceanothus. A similar application made on September 14, 1962 showed only 74% mortality of the gall larvae. This lower kill may have been due to the greater proportion of older, possibly more resistant, larvae at this time than at the June 27th spraying. A soil drench of the dimethoate, made at six times the strength of the foliar spray, applied around the base of Ceanothus shrubs in Pasa- dena on July 11, 1962, resulted in 70% mortality of the gall larvae, but 196} GALL MOTH 187 M E T R I C .j_,. J_j_j. 1 1, i S Y S T E .M Fig. 1. The adult moth, Periploca ceanothiella (Cosens), at bottom right; larva and pupa at left; larva in cut gall at top. 188 MUNRO /. Res. Lepid. I> i(}):iS}-i90, 196} GALL MOTH 189 Fig. 3. Stem galls from which the moths have emerged; note the exit hole in the encircled gall. The twig at the right has been completely killed by the larvae. Note that the galls occur at the beginning of each period of stem growth, and that in the twig at the left, two different years’ galls are shown. 190 MUNRO /. Kcs. Lcpid. caused burning and subsequent defoliation wherever the drench came in contact with the low-growing foliage. Examination of galls from untreated Ceanothus coinciding with observation on the treated shrubs, showed natural mortality of less than one percent, and of no significance. Limited observations on Ceanothus growing in the wild showed the stem-galls to be less abundant than on Ceanothus grown under culti- vation. This might indicate that irrigation and other care which these ornamentals received under cultivation makes them more attractive to this gall insect. SUMMARY The stem-gall moth, Periploca ceanothiella ( Cosens ) , is a pest of ornamental shrubs in the genus Ceanothus. The moths emerge from the galls and lay their eggs throughout spring and early summer. There is but one generation or annual brood. Three naturally occurring ichneu- monid parasites accounted for up to 19% mortality of the maturing gall larvae. No hyperparasites were observed. Examination of 40 species and varieties of Ceanothus growing in mixed collections at several southern California locations showed 2 with heavy occurrence of the galls, 4 with moderate occurrence, 13 with light occurrence, and 21 with no occurrence. This would suggest the planting of the more non-susceptible varieties to alleviate or avoid the problem where practicable. Dimethoate used as a foliar spray proved highly effective against the immature gall larvae, as well as the leaf miner, thrips, and mites, but had no apparent effect on the mature gall larvae and pupae, or the parasites. Limited observations indicated the galls to be more troublesome on Ceanothus growing under cultivation than in the wild. REFERENCES CITED VAN RENSSELAR, M., and H. E. McMINN. 1942. Ceanothus. Gillick Press, Berkeley, California. 308 pp. MUESEBECK, C.