DETERMINATE EVOLUTION IN THE COLOR- PATTERN OF THE LADY-BEETLES BY ROSWELL H. JOHNSON WASHINGTON, D. C. PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON 1910 CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION No. 122 PAPERS OF THE STATION FOR EXPERIMENTAL EVOLUTION No. 15 Copies of this Book were first iss JUN291910 THE CORN M \N I'ltl NTI N<; COMPANY, ( \ i: i isi. I-:. i'i:\ \>i i.\ \ \ I A. CONTENTS. Page PART 1- INTRODUCTORY CONSIDERATIONS Principles of classification 4 Structure of pigmented areas 6 Development of pigmented areas Utility of the color-pattern 9 Life and habits 10 Variation and heredity of body-length 13 Acknowledgments 14 PART II.— THE DATA ARRANGED BY SPECIES 15 Hippodamini 15 Anisosticta 15 strigata 15 Naemia 15 seriata 15 Nacronaemia 16 episcopalis 16 Paranaemia 16 vittigera 16 Megilla 16 maculata 16 Ceratomegilla 17 ulkei 17 Eriopsis 17 connexa 17 Hippodamia 1" glacialis 19 convergens 21 bowditchi 45 spuria 46 oregonensis 48 cockerelli 49 dispar 50 sinuata 50 tredecimpunctata 50 americana 52 parenthesis 52 apicalis 54 lengi 54 falcigera 55 Coccinellini 56 Neoharmonia 56 venusta 56 notulata 56 ampla 56 iii iv CONTENTS. PART II. — THE DATA ARRANGED BY SPECIES— Continued. Coccinellini— Continued. Paue Coccinella 57 perplexa 57 tricuspis 59 novemnotata 5'.» johnsoni 61 transversoguttata 61 calif ornica 62 monticola 63 difficilis 64 suturallis 64 prolongata' 64 Cycloneda 65 sanguinea 65 munda 65 ater 65 Olla 66 abdominalis 66 pi agi a ta 66 Adalia 67 bipunctata 67 frigida 68 annectans 70 Cleis 72 picta 72 hudsonica 72 Agrabia 73 cy anoptera 73 Anisocalvia 73 duodecimmaculata 73 quatuordecimguttata 73 Anatis 74 quindecimpunctata 74 mali 74 rathvoni 74 lecon tei 74 Epilachnini 75 Epilachina 75 borealis 75 toweri 78 corrupta 79 mexicana 79 PART III. —GENERAL DISCUSSION 81 Variation 81 Modification 84 Distribution 86 Heredity 91 Phylogeny 94 Evolution 95 Summary of conclusions 102 Bibliography 103 DETERMINATE EVOLUTION IN THE COLOR- PATTERN OF THE LADY-BEETLES BY ROSWELL H. JOHNSON PART I. INTRODUCTORY CONSIDERATIONS. The lady-beetles were considered desirable for this study because their variability, distribution, and taxonomy indicated that they had been recently and are probably now in an active state of evolution. The prac- tical advantage that they could be obtained in considerable numbers from many localities and could be experimentally bred was, then, decisive in their favor. All of the American coccinellids which could be obtained in numbers and which showed a variable color-pattern were studied, viz, Hippodamini, Coccinellce, and Epilachnini, divisions employed by Casey (1899). The Epilachnini differ from the rest of the family in being leaf-eating, in having longer generations, and in hibernating necessarily in the winter; whereas the other coccinellids which eat aphids, fungus spores, and pollen will remain active during the winter in a vivarium and give many more generations. The difficulties of keeping a large and constant stock of aphids on hand and keeping the beetles free from diseases, especially such as result from dampness in the late summer and irregular temperatures in the winter, proved to be very serious and prematurely shortened many pedigrees. These difficulties also decreased the numbers which could be successfully managed. Nearly all the pedigrees, however, are given, for when too fragmentary to have much value in the study of dominance and segregation, they are, nevertheless, of value in showing the transmissi- bility of some characters and variation from the parent. The dorsal color-pattern in all the American species within these groups is given for the sake of completeness, although for some species but little variation data was obtainable. This fact, together with the necessity of carrying in mind the data of variation and distribution in discussing experiments upon any given species, has led me to arrange the presenta- tion according to species, after this introductory section. Furthermore, this arrangement will be more advantageous for those whose interest in this contribution is primarily taxonomic. Every considerable new accession of data involves another revision of these tribes, since there has been so much disagreement among the authors and since the lines between varieties and species in these remarkable genera are so doubtful. I am obliged, therefore, to present a revision of my own in order to have a suitable nomenclature. Since the revision is not an end in itself, I have not given full descriptions, but only discussed and illustrated the color-pattern of the parts studied, although of course the revision is based upon many characteristics. Not having studied the structural features of the foreign species, I have not ventured to revise the genera and have simply adopted the genera as used by Casey. EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. The differences in sculpture of the elytra have not been found to have the constancy and systematic value which some of the earlier writers have placed upon them. The variation is, moreover, largely ontogenetic, as I have observed in the pedigreed material. The shape of the elytra must also be used with considerable caution, because of its variability, as shown in table 2, and the optical illusions from the angle at which they harden and from the longitudinality or the reverse of the pattern. The sex of the individuals has in most cases been observed; always in the experiments. The same sex is not the more heavily pigmented in every species. The sexes are not given separately except where the difference was obviously significant. The color-patterns of the larvae have been used in evaluating relationships in some cases, but their description is outside the field of this paper. The numbering of the elytral spots employed in Hippodamia coiiveru<'iix follows Weise, not Kellogg and Bell (see figs. 3, b, and 17) . In other genera a corresponding system is employed: + is used to indicate confluence, apto indicate a close approach to confluence, s/+ for a very slight confluence, and tr for transition between separateness and confluence. Formulae are used where possible to designate unnamed varieties; in other cases it has been necessary to use letters. New names have been given to varieties only when they are common, distinct, and frequently collected, so that a name is needed. PRINCIPLES OF CLASSIFICATION. There is an improper demand for finality of treatment on the part of many systematists. These workers wish to use a treatise almost solely for the purpose of definitely dividing their collected specimens into categories. Such a demand leads many systematic writers to a false positiveness in regard to distinctions and evaluations. Many revisions are thus led to present a precision and show of finality which is not justified. New names involving new distinctions are valuable, but the reviser who fails to bring to light the points of doubt and uncertainty as well is negligent. A revi- sion should give questions as well as answers; it should not only shed light, but indicate where more light is needed. To assume surety on an inadequate basis is inimical to the advance of science. On the other hand, there are some who go to the other extreme and hesitate about calling attention to an apparently new species, or some other difference, until they feel wholly assured. This policy retards progress and may result in the information never being made public because of the intervening death, incapacity, or disinclination of the investigator. An impression based upon extensive special knowledge has scientific value, provided it is avow- edly only an impression. I have not sought, therefore, in this revision any degree of finality, but have tried to adopt a nomenclature which, from PRINCIPLES OF CLASSIFICATION. 5 present data, seems the nearest approximation I can make to the true condition, for I well know that its usefulness will soon be past. I believe that it is very important that the International Congress of Zoologists should adopt a set of definitions of species and the less inclusive taxonomic grades in order that some uniformity can be attained in the use of such words as varieties, subspecies, forms, aberrations, mutations, etc. Until that time each writer must define his use of such terms when- ever called upon to use them extensively. Failure to do this is causing a constantly growing confusion. The following definitions will show the meaning of the terms as here employed. I am well aware that many of these terms are used differently by some writers, but at present each worker can merely choose the use that seems most satisfactory to him. A species consists of individuals which, aside from sex, age, and periodic differences, have a fundamental similarity and which are habitually and successfully interbreeding. They comprise an intergenerating unit. A section is a part of a species which is cut off by some barrier from intergeneration. Where the individuals of the section have no apparent difference from the mass of the species the section is ignored. Where the difference is so great that individuals of the two sections are always readily distinguished, a distinctive name becomes desirable and the form is considered another species rather than a section, although experiment- ally the two will interbreed freely. The difference between section and species is one only of degree of difference in the characteristics. The sec- tion may be found to be ontogenetic or phylogenetic in its nature. A subspecies is a part of a species which inhabits only a portion of the range of the species and which, while differing in some appreciable respect, intergrades with it in the intermediate region. At any one point through- out the range of the species the quantitative expression of the chief differ- ential of a number of individuals similar in age, sex, and season should give a unimodal polygon of frequency. The term "subspecies" maybe used without reference to the inheritability of the differences. But where these differences are found primarily because they are the result of environment in each generation, the subspecies is called an ontogenetic subspecies. Where it is present primarily because inherited it is called a phylogenetic subspecies. If the conditions of a subspecies are fulfilled, except that the individuals in question are scattered geographically and are confined to a particular kind of habitat throughout its range, then we have a habitat-form, which may be ontogenetic or phylogenetic. The usual assumption that all habitat-forms are ontogenetic is, I think, questionable, although they are probably generally so. The distinction between a subspecies and a habitat- form is not a sharp one, for a subspecies is a habitat-form where the habitat involves a very large, continuous area. The habitat-form falls 6 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. short of a subspecies when the habitat is made up of many discontinuous areas surrounded by a population of the contrasted forms. A variety is a part of the species which differs from the typical members of the species by some hiatus which can be detected by the decreased number or absence of intergrades. When some considerable imperfection of the interfertility of the apparent variety and the typical species is found or some other cause makes their interbreeding- uncommon, the variety becomes a species. Only an arbitrary line can be drawn, and to make an arbitrarily definite one would be infeasible at the present stage of our science. The variety may be equally common throughout the range of the species; it may be much more common in one part; it may be limited to one part of the range; or it may wholly replace the species in one part of the range. If its frequency only gradually changes in one direction, a subspecies is there constituted. If the variety is confined to a particular habitat, we have a habitat-variety. An aberration is an individual which is wholly unique, generally patho- logical in origin, and often unsymmetrical. Where the character is sym- metrical and there are no pathological appearances or the variation is of a sort found normally in other species, the specimen is probably one of a rare variety rather than an aberration. A mutation is a specimen of a variety which is known to be capable of hereditary transmission, but the ancestors of which are typical individuals of the species. Where this information is lacking the apparent mutation can only be considered to be a rare variety, the origin of which is probably traceable to a progenitor more or less remote. An extreme fluctuant is an individual which appears sufficiently dif- ferent to be noticeable, but is, nevertheless, only an extreme case of the ordinary fluctuation, as is shown by the greater frequency of the inter- mediate conditions. Form is a term of convenience only, applied to differences the real nature of which is in question. Many forms will later be found to be extreme fluctuants, others may represent points in a fluctuating series easily recognized or described, while still others are doubtless real posi- tions of organic stability, which further study will show to be varieties or intermediate conditions between varieties and continuous variations. STRUCTURE OF PIGMENTED AREAS. I have limited the work, with a few exceptions, to the color-pattern, be- cause with the other characters we have for the most part stable conditions with only a little fluctuating variation. For the further advantages of specialization I have concentrated my attention upon the pronotum and elytron. I believe, however, that similar results would have resulted were the study extended to include the coloration of the larva and the remaining parts of the imago. The coloration of the pupa, however, is rquch more subject to modification. STRUCTURE OF PIGMENTED AREAS. That the attachments of the muscles have an influence on the color- pattern of the head and pronotum is probable. I have not felt it neces- sary to go into the matter of structure of these parts in detail, however, because the great range of variation in the color-pattern within a genus, or even within a species, where the structure cowld not differ in any but a slight degree, is such that the influence of structure can be only along narrow lines. In the elytra the influence of structure is more evident, but less than might be expected. The sutural margin is thickened and paralleled by a slight groove. As will be seen later, this margin is especially subject to pig- mentation, and this in fact distinguishes several species and varieties. Spots are confluent with it in a few species and may become so by modifi- cation by cold in Hippodamia cowwrfieus. Yet, in general, when spots near the suture enlarge, their margins, which if they remained circular would become tangent to the suture, flatten and become parallel to the suture without touching it. Fie. 1.— Position of linea interim, liix-a nif, f'nci'iiiellit: c, ffiiipotlainia: il, Anulix; e, h'/ii/ac/inn. The lateral and basal margin is much modified structurally in variations of germinal origin. It is one of the parts least subject to pigmentation. In modification by cold, however, pigment readily extends along it from neighboring spots. In Adah' a three veins are visible, named by Schroeder (1901) linea interna, linea media, and linea externa (fig. 1). The linea externa is quite close to the reflexed margin. It departs farthest at the humeral angle, where it can be most clearly seen. Any other veins that exist must be in the modified and thickened sutural or marginal parts of the elytra. The influence of these lines on the color-pattern of Adalia is considerable, as shown by Schroeder (1901). But in Coccinella the linea externa can not be seen in the hardened ely- tron; yet its path can be traced in the basal half of the elytra by the extension of pigment along its course in C. iwvemnotata from the spots 1 and 2 along the linea externa. This is farther from the margin than in Adalia. It is probable that there is a vein in the soft elytron, visible evi- dence of which (to the hand-lens) does not persist in the mature wing. In Hippodamia also the linea externa is not visible. Its course is outT lined by pigment in cases of modification by cold and in the two wild EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. specimens shown in figs. 25, 16, aa, and 16, 66. In Epilachna and Anatis it is not visible and no cases were seen of pigment disturbance. Pigment follows along the veins readily, especially along the linea ex- terna, in cases of increase of pigment by modification; but in varieties of spotted species of germinal constitution the influence of veins is not great. The vittas wrhen present are frequently at an angle to the course of the veins or broader upon one side of it than the other. Transverse confluence, which is not influenced by any elytral structures, is more common than vittae. There is pressure upon the elytron from below at the humeral spot by the basal attachment of the wing below, at the site of spot e by a projec- tion of the thoracic plates, at and near the site of spots 4 and 5 by the fold- ing of the wings beneath, and at spot 6 by the tip of the abdomen when bent back, as it frequently is. Nevertheless these spots are not infre- quently absent and there is no especial pressure where other spots are present. From these considerations and the well-known theoretical objections to kinetogenetic origins of germinal characteristics, I am not disposed to see any causal relation between pressure from below and the position of the spots, although the coincidence is striking in the second case. DEVELOPMENT OF PIGMENTED AREAS. The color-pattern of the head and thorax are fully formed at time of emergence. The black pattern of the elytra is formed subsequently. The soft, freshly expanded elytra are free from all black pigment. This is true in all the species which I saw emerge, including Epilachna borealis, of which the contrary has been stated. No sequence in the appearance of spots could be positively determined, so nearly simultaneous is their development and so gradual is their appear- ance. Where the pattern is composed of strongly confluent spots, the pattern appeared at once, though not at first of full intensity, without pass- ing through a separate spot stage. The pigment area, however, increased slightly during the few successive days after emergence before the elytra were fully hardened. This extension was so slight that it would not have been detected but that the pattern of a few freshly-emerged beetles, the pigment of which had become fully black, were described as having spots 1 and 3 very close together, but separate. Several days later they were found to be slightly confluent. Confluence of this kind, however, is only seen where the contiguous spots are unduly large and is the result of a mere enlargement of spots. Their inheritance is like that of an extreme fluctuant, as shown in pedigree table 15. Confluence, therefore, is of two kinds. In confluence of the hereditary type, the spots may be quite small and distant, as in Epilachna corrupta. THE UTILITY OF THE COLOR-PATTERN. 9 THE UTILITY OF THE COLOR-PATTERN. The utility of the color-pattern is obviously not that of protective resem- blance, for it is decidedly conspicuous in the green surroundings which the beetles frequent. This conspicuousness is generally thought to be expli- cable as a result of the habit these beetles have when disturbed of expel- ling a disagreeable fluid from the ends of the femora. The possession of a striking and peculiar appearance, by which the bird or other pre- daceous animal might associate this painful experience, would obviously be advantageous. In Poulton's terminology the color is aposematic. The only other explanation that seems at all plausible is that of sexual attrac- tiveness, i. e., the color is epigamic. The apparent indifference with which Hippodamia convergens, spotted or unspotted, breed together at Berkeley, California, although both forms are common there, makes this unlikely. I collected a number of mating pairs and could see no evidence for either assortative or preferential mating, in this confirming Kellogg, who also collected a number of mating pairs and reached the same conclusion. The explanation of the ground-color of these beetles as warning or apo- sematic may be accepted upon the following evidence from Judd (1899), based upon extensive examinations of the stomachs of birds: In the Coccinellidas we have showy insects, ill-scented or flavored, that are eaten by but very few birds— the flycatchers and swallows ; and hence here is a whole family which conforms well with the theory of warning coloration. * * * The showy, ill- flavored Coccinellidae [are] * * * almost as * * * highly protected from birds as the hairy caterpillars and the elm leaf-beetle. To this may be added my attempt to feed Coccinella novemnotata to a catbird reared in captivity. The bird took the beetle up to its perch, but then dropped it. The beetle walked away without further molestation. Since Judd finds that the barred and spotless species as well as the spotted species are protected, the particular patterns can be of little im- portance, although they are so diverse. In flight, where the beetles are most attacked, the pattern is inconspicuous, while the ground-color is evi- dent. But what is quite conclusive is the origin and persistence of varieties having fundamental differences from the pattern of the spots of the par- ent species. The fundamental usefulness of an aposematic pattern would depend, above all, upon its constancy. If the spotted pattern, which is so widespread in the family and is the primitive pattern for many genera or subgenera, has a strong association value with distastefulness, all departures from it would be suicidal; yet we find they are not. It is certain, then, that there is no high selective value in the spotted pattern, but it is possible that there is a feeble one. The loss of the spots or their change into other patterns must, then, be the result of some stronger evolutionary force, which I believe to be determinate evolution, over- coming the slight advantage placed on spottedness by natural selection. 10 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. This cause will be discussed after the presentation of the data. It is referred to here that the reader may bear in mind the question of utility cerans determinate evolution as the species and varieties are described. LIFE AND HABITS. It is not necessary to give any detailed account of the habits of these insects for the purpose of this article. Only those features will be selected that are significant for this discussion of the color-evolution. Lady-beetles are probably well protected from predaceous animals by their distasteful- ness, as shown by the experiment with the catbird. In the localities where they are abundant I have never seen them attacked. Frequent causes of death are parasitism by internal insect larva?, an unknown dis- ease, and difficulty in casting the larval skins on emergence from the pupa ; but the two great causes of death are hibernation and the maldistribution of eggs. Any considerable advantage in either of these respects would be strongly favored by natural selection. Epilarl/tnt is uncommon in the spring at Cold Spring Harbor until the new brood comes forth, when it is quite abundant. Efforts to hibernate it under as favorable conditions as I could arrange succeeded in not more than 50 per cent of the individuals. Attempted hibernation with other species was unsuccessful, although some, when provided with food, overlived the winter in the vivarium. The critical conditions seem to be extremes of dryness and moisture. MeyiUo nmctiJtita hibernates, at least frequently, in masses, one of which was found and kindly sent me by Dr. Robert W. Hall, of Lehigh University. I am assured by Prof. N. F. Davis, of Bucknell College, that in the spring such a mass was seen to mount a fence-post preparatory to flight. In the Western States Hippodamia con- vergent* and spuria.Yesort to the same practice. Strangely enough, such masses were frequently found on mountain-tops throughout the Western States. This is probably not adaptive, but a by-product of some tropisms. These beetles are found in great numbers in the flotsam of the shore of large bodies of water when a certain sequence of winds occurs during the time when large numbers are in their long flight. The ability to collect large numbers on mountain-tops and shores has been a favoring circum- stance to the collection of material for this paper1. I would like here to solicit the opportunity of examining any such masses, and I will gladly return them if desired. While these beetles fly very little in cloudy weather, on warm sunny days they frequently take long flights. This is important as breaking down the probability of isolation, widening the range of varieties, and making their passage into species more difficult. Dispersal is probably even more effective than in birds, which have such powerful homing instincts, although of course the occasional storm-driven bird frequently gets much farther astray. LIFE AND HABITS. 11 The food of these beetles is well known to consist principally of aphids. Certain coccids are eaten by some of the species, but this is the exception. A coccid of the chestnut constitutes one of the principal foods of Cijcloneda munda at Cold Spring- Harbor. This coccid is not eaten by any of the other species. Meyilla fiiscilabris is also peculiar in its food habits, for it eats a much larger proportion of pollen and fungus spores than are eaten by the other species. In general, coccinellids eat a wide range of species of aphids. Some species of aphids are found to be especially attacked by some species of lady-beetles; thus, that of Rhanumx c'imiles. FIG. 2.— Variation ol' length ol'rl.vtra of JIij>i>ml;i; Sand Hill;;. Nebraska. Thep-e specimens piobably represent a subspecies, for a specimen showing nearly as much separation came from Beaver Dam, Wisconsin. Specimens of the typical species of the variety (lohrifinti are rare. Var. B. A specimen from Wyoming shows spots fused into two marks, except the juxta- sutural spots, which are free. Weise records 2 specimens of A. novemdecimpunctata var. irrcniilaris with 1, 2, 3 -I i, 4 i 5, 6 the suture, 7 ! 8, 9 (his numbering), from Oregon. This throws some doubt on the specific distinction between A. novemdecimpunctata and stririata. The difference in the elytral markings are certainly bridged over by our variations. Genus NAEMIA Mulsant. 3034. Nacmia seriata Mulsant. Distribution: Maritime regions of the eastern United States and south- ern California. Form n. The spots separate, but the apical spot meets the suture. Form hf Some of the spots separate, others united. Mulsant says the separation of the third spot from the ground is most frequent, followed by the separation of the third from the fourth. Form c. Scutellar mark prolonged and extended to the second spot. Snh^/if cirs <>f Florida: Pigment reduced, so that the spots are not confluent. In this it is analogous to the subspecies floridana of McgiUfi macuhita. Pronotal pattern resolved to 4 spots in some cases. Subspecies litiginosa Mulsant. Specimens from Southern California in the Leng collection have the spots less con- fluent longitudinally, but more so transversely. They also have the caudal white spot on pronotum larger. 15 16 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES- Genus MACRONAEMIA Casey. 3035. Macronaeraia episcopalis Kirby. Distribution: Western States and Canada. The pronotal pattern in a specimen from Assiniboin, Montana, and in another from Wyoming is reduced to 6 spots. Otherwise it is relatively constant. Genus PARANAEMIA Casey. 3037. Paranaemia vlttigera Mannerheim. Distribution: Colorado, New Mexico, Arizona, California, and New Mexico. Shows very little significant variation. Casey believes those of Colorado and Arizona to be more slender, but with similar coloration, and has called them P. h I.'K,. :;.— Variation of the elytral pattern of Meg ilia macidata. c^ mode; a to y from Shawnee, ' Pennsylvania; h from Ohio and Texas. Genus MEGILLA Mulsant. 3036. Megilla maculata DeGeer. (Figs. 3, 4.) Distribution: United States and Canada, except Pacific coast. This species is more constant than the other spotted Hippodamini (fig. 3), as indicated by the following count of 321 individuals from Shawnee, Pennsylvania. Normal, 310. Form a. The elytral spot at 3/4 on the suture resolved to a spot on each elytron free from the suture; 4 individuals. Form b. 1 + scutellar mark; 2 individuals (fig. 3,/.) Form c. An extra spot between 1, 2, and the scutellar mark; 1 individual (fig. 3, c/) ; 1 also seen at Cold Spring Harbor, New York. Form d. 3 + 5; 1 individual (fig. 3, g) ; I also from Texas. Form c. Spot 2 resolved into a larger internal and a smaller lateral spot (fig. 3, fe). There was a case of a slight deposition of pigment along the vein between 1 and 2 and in another between 4 and 6, in each case upon the right side only. Also one case of 2 + 3 on left side only. Form/. Scutellar mark + 2. Not taken in the Shawnee lot, but 1 seen from Texas and 1 from Ohio (fig. 3, h). The typical species is found only in Cuba. It has but one continuous area of black pigment upon the pronotum. The two subspecies following may be entitled to specific rank. The decision must wait for more ex- tended collection in the Southeastern States. Fi<;. -4.— Variation ol' the pronotal pattern of ^let/ilia maculti/ti. b mode; n to c, from Shawnee, Pennsylvania. CERATOMEGILLA— ERIOPSIS- HIPPODAMIA. 17 Subspecies fuscilabris Casey. The pronotum pattern has the two areas not confluent (fig. 4) and the spots not greatly reduced. This characterizes all of the North American range except the south- eastern States and Cuba. Casey believes that the specimens from South Texas are broader, with the color-pattern the same, and has named it Megilla strenua. Subspecies floridana Leng. Specimens from Beaufort, North Carolina, to Louisiana are smaller. The pigment much reduced, but the pronotum, although having the pigment reduced often to 4 spots may in other cases show a transverse confluence, in this respect approaching the typical species as found in Cuba. Subjection to both 40° C. and 18° C. in the usual way failed to produce significant modification. Genus CERATOMEGILLA Crotch. 3038. Ceratomegilla ulkei Crotch. (Fig. 5.) Distribution: Hudson Bay. A monotypic genus having unique antennae and of obscure relationship. Apparently none have been taken since Ulke found the type at Hudson Bay. The drawing given (fig. 5) is pinned into the Leconte collection in the Museum of Com- parative Zoology. 5 Genus ERIOPSIS Mulsant. 3040. Eriopsis connexa Germar. Distribution: Texas, California, and Vancou- ver Island. n n * • • i /» -i • , -i FIOJ. •>.— Elytra 1 pattern in A South American species rarely found in the ccratomeguin United States (fig. 6). It is possibly derived Fl<'j from a spotted Hippodamia by reversal of pat- tern. E. eschscholzii from Chili represents an intermediate condition com- parable to Hippodamia cockerelli. Genus HIPPODAMIA Mulsant. The species of Hippodamia fall into several distinct phylogenetic sections, as shown in fig. 7 and table 1. The comparative size and form of some of the species of this genus are shown in table 2. While these differ, the ranges overlap to such an extent that size and shape are seldom service- able in the identification of single specimens. The differences are large enough, however, so that the eye readily detects them in the comparison of series. The females are uniformly larger, but differ little in propor- tions from the males. Under H. convergens similar data will be given in regard to two of its varieties. 18 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. GLACIAUS 80WDITCHI ' } CONVERGENS DISPAR COCKERELLI OREGONENSIS SINUATA SPURIA TREDECIM.PUNCTATA AMERICANA PARENTHESIS APICALIS LENQI FALCIGERA FIG. 7.— The species of Hii>podaini 29 33.9 37 l*.5 53.8 57.5 Female 75 54 68.2 73 21 26.4 31 82.5 38.7 44.5 29 35.5 39 47.5 52.1 56.5 H. parenthesis, Stony Lake: Male 53 56 59.9 61 19 24.2 29 32.5 40.5 46.5 30 33.7 37 51.5 56.3 60.5 Female 85 53 62.8 68 20 25.0 30 32.5 40.0 47.5 30 33.8 38 50.5 51.1 61.5 3044. Hippodamia glacialis Fabricius. Distribution: United States and Canada. The color-pattern (fig. 8) of this species is approached and even realized in some cases by specimens of H. convergens in the Western States beyond the usually recognized range of H. glacialis. Yet complete intersterility was found to prevail in repeated tests between eastern specimens of the two species. Whether this intersterility is bridged over by these speci- mens in the Western States I was unable to test. 20 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. The resemblance of this species to H. convergens, and especially the simi- larity of their larvae, is such that it has probably been derived from that older type at no very distant date. The pronotal pattern is like that of H. convergens, but shows relatively less variation (fig. 9). Var. A. The humeral spot absent. In 9 out of 26, or 35 per cent of the males, and in 6 out of 29, or 21 per cent of the females at Cold Spring Harbor. Var. B. Spots 4 + 5 + 6. Several taken from di- verse localities. Nowhere established (fig. 8, /). Var. C. Spots 6+4 + 5. As above (fig. 8, /?). Var. D. Spots 4, 5, ; nd 6 merged in one rounded area (fig. 8, ./). Only 1 specimen seen. Var. E. With spot No. 2; 3 cases in 55, or 6 per cent, at Cold Spring Harbor (fig. 8, a). Var. F. With the extensa mark. Only 1 specimen at Cold Spring Harbor (fig. 8, sr). Var. G. Spots 4, 5, 6 (fig. 8, rf). Var. H. Without discal spots on pronotum. Only 1 specimen from Cold Spring Harbor. cialis of the Western States. Var. /. Median white spot upon head extending to eyes, western specimens (fig. 10). FIG. s. — Variation of elytra! pattern of ffippodamia b = mode. Also seen in the H. convergens var. Especially characteristic of 3T d FIG. 9. — Variation of proiiotal pattern of IIii>podatnin ijlncnilis. it = mode. Fi<;. 10. — Variation of color-pattern of head of Hippodamia i/lncinlix. d = mode; K from Omura, Kansas; c to e from Cold Spring Harbor, New York. The appearance of spots 2 and 3 and of the scutellar mark, which is occasional, may be looked upon as reversionary. The extensa mark is also met with in H. convergens var. extensa and is a case of parallel variation. HIPPODAMIA. 21 HEREDITY. The three females in table 3 with unknown mates show clearly enough that the presence or absence of spots 1 and 2 is inheritable. No. 453 had the transverse band broader than usual and its progeny show the inherit- ance of this condition, for in some of the offspring- the mother was exceeded in this respect and none of them show the slightest tendency to the separa- tion of the component parts. The heredity of spots 1 and 2 is segregative, with a few intergrades. The shortness of the pedigrees leaves the ques- tion of dominance unsettled. TABLE 3.— Heredity in Hippodamia glacialis. Mother. No. 1 •2 1 and 5 No. Pattern. Present. Very small. Absent. Present. Absent. Sepa- rate. I'nited. 248s 1 absent, 2 present, 4 + 5 S 3 (1 5 7 1 3 5 434 1 present, '_' absent, 4 + 5 31 L>5 i> :; 1) 31 0 :;i 453s 1 present, 12 absent, •1+5 14 13 0 1 0 14 0 14 3046. Hippodamia convergent Guerin. Distribution: North America. This is a wide-ranging species which is highly variable. For reasons given later I have thought it best to reduce to the status of varieties several of its derivatives which have received specific names. The variation (fig. 11) is strikingly parallel to that in Hippodamia septempunctata and to a less degree to that in Adonia varieyata. The varieties of these two Euro- pean species have been well studied and named. This is not the case with the American H. convergens. I have indicated the correspondence of these varieties by the sign of equivalence ( - ) used by mathematicians. The correlation of pronotal and elytral patterns is low enough to make it desir- able to treat them separately. Types of Patterns in Elytral Spots. Spots £, 1, 2, 3, 4, 5, 6. Typical. Spots A +3 + 1, 4 + 5, 6, va>-. quinquesignata Kirby. Spots J +3 + 1, 4 + 5, 6, with pronotal discal spots, var. puncticollis Casey. Spots i -|- 3 -|- 1, 2, 4 -f- 5, 6, H. septemmaculata var. continue/,. Spots | + 3 + 1, 2, 4, 5, 6. Spots £ -f-3, 1, 2, 4 + 5, 6, var. caseyi (new variety) - >A. variegatavar. ustiilata Weise. O H. tredecimpunctata var. contorta Weise. This common variety is generally called lecontei Mulsant, the description of which calls for a pattern quite different, which is given below. This variety is so well known that it seems best to rename it. Var. dejecta (new variety). Formula as in var. caseyi, but spot 1 small and 1 -(- 3 much less heavily pigmented. Because of its interesting relation with var. caseyi, discussed later, this variety is given a name. EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. FIG. 11. — Variation of elytral pattern in Hippodamia conreryens. HIPPODAMIA. 23 Spots Spots Spots Spots Spots, Spots Spots Spots Spots Spots Spots Spots Spots \ + 3, 1, 4 + 5, 6.0 H. septemmaculata var. viadri Weise. J + 3, 1, 2, 4, 5, 6, = = H. septemmaculata var. puyknlli Weise - H. tredecim- punctata var. triloba Weise. & + 3, 1, 4, 5, 6, =c= H. septemmaculata var. oblonga Weise. i, 1, 2, 3, 4,5 + 6, =0= H. tredecimpunctata var. c. nigrum; rare; California and Nevada. 4, 1, 2, 3, 4, 5 + 6, var. juncta Casey; taken in Sonoma County, California, and 1 specimen from Stony Lake, Michigan. i, 1, 2, 3, 4+ 6, 5, rare; Pacific States. i, -)- 3 + 1, 2, 4 -) 5 -j- 6, rare; Rocky Mountains. £, 1,2, 3, 4 + 5, 6 z H. tredecimpunctata var. spissa Weise =:=#. septemmacu- lata var. aestiba; common in Western States. J, 1, 4 + 5, (>, pseudoglacialis (new variety); New Mexico and northward. J + 3 + e, extensa Mulsant, Western States. i + 3 + e, 2, 4, 5, 6. £ + 3 + e, 2, 4 + 5, 6. J + 3, e, 2, 4, 5, 6. FIG. 12 — Variation of elytral pattern of the quindecimmaculata varieties. c = mode; «, from Lake Superior; /*, California; c, St. Louis. Missouri; , quindecimmaculata A Mulsant; very rare; Central States. Spots \, 1, 2, 3 + q, 4 + 5, 6, quindecimmaculata D Mulsant; Missouri Valley; very rare. Spots J, 1, 2, J + 3 + q, 4, 5, 6, quindecimmaculata C Mulsant; Missouri Valley; very rare. Spots \ + 3 + q, 4, 5; 1 specimen from Fairfield, Washington. Spots i, 1, 2 + q, 3, 4, 5, 6; very rare; 1 specimen from Kamiack Butte, Washington, upon one side only. Spots £, 1, 2 + q + 3 + \, 4, 5, 6; very rare; Missouri Valley. Spots \, 1 + q + 3, 2, 4 + 5, 6; 2 specimens from Keeler, California, and St. Louis, Mis- souri, respectively. With the comma-mark as in Neoharmonia venusta var. A. Spots 1, 2, 3, 4, 5=c= septemmaculata var. vorax Weise. Eastern as well as Western States. Several spots lacking, convergens C Mulsant; Western States. Mark \ only, obsoleta Leconte; Pacific States and Western Mexico. Most numerous in western Oregon outside the humid region. Extends in reduced numbers through Arizona to the Great Plains, giving way to var. C Mulsant gradually. Spotless, with obsoleta; rare; = H. tredecimpunctata var. signata Paid. = ^= H. septem- maculata var. r libra Weise. 24 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. Var. moesta Leconte: The black area extended so as to leave only a small basal lunette and a small area between 4 and 6 of the reddish ground-color. Northern California to Vancouver Island and British Columbia between Hope and Okanogan. Limited to the humid Pacific Coast region. I have been unable to obtain living specimens. There is reasonable doubt as to whether this pattern may not constitute a distinct species, especially since at Dilley, Ore- gon, where the spotless variety prevails, one has been taken. It is clearly a melanic derivative from a spotted Hippodamia. Two specimens with pigment reduced enough to reveal the unobscured pattern are shown in fig. 13. While they do not agree in showing one direct line of development, it is probable that it is through var. quinquesignata. It is likely, moreover, that the beetles have some such pattern when newly emerged from the pupa, as Miss Isabel McCracken has shown for the melanic variety of Melasoma scripta, and that the variation in these when known will make the ancestry of var. moesta clear. The modification experiments uponH. convergens prove that a cold environment in- creases pigmentation. The degree of cold was such, however, that the beetles suffered more or less damage, and still they fell far short of the melanism of the variety moesta. Moisture alone produced no effect. For this reason and because Dilley, Oregon, has typical var. moesta rarely along with the spotless H. convergens, I believe that H. moesta is an inherited condition and not an "ontogenetic species." It remains to be seen whether it belongs to the intergenerating unit of H. convergens or not. FIG. i:!.— Variation of color-pattern of var. moesta. with twointergrades. e =-- mode ; b, from Humboldt County, California. Examination of these varieties shows that they are the several combina- tions of a number of characteristics, as follows: 1 to 7. The mark £ and the spots 1, 2, 11. The confluence of 1 and 3. 3, 4, 5, 6, respectively. 12. The confluence of 4 and 5. 8. The new spot e. 13. The confluence of 4 and 6. 9. The new spot q. 14. The confluence of 5 and 6. 10. The confluence of £ and 3. 15. The melanism of var. moesta. The combinations of these characteristics produce the different vari- eties given and the numerous ones not given, characterized by loss of certain spots. The distribution of some of the characteristics is given in tables 4 and 5. The pronotum is subject to a similar remarkable variation (fig. 14), and this, strangely enough, is not highly correlated with the coloration of the elytra. Spotless elytra, as well as the typical spotted pattern, are asso- ciated with all or nearly all of the pronotal patterns. The typical pronotum is black, with 2 light convergent discal marks and a narrow light margin. The black area is slightly emarginate mesocephalad and the white margin is slightly narrower laterad and broader cephalolaterad and caudolaterad. The convergent marks may, on the one hand, disappear, or, on the other hand, break through to the margin, generally at a point in the continuation HIPPODAMIA. 25 of the direction of its greatest length. Occasionally it opens caudad of the lateral process and at other times at both places. Still more rarely it breaks through at a third point more directly cephalad. Table 5 shows the geographical distribution of these characteristics and its association with the elytral pattern. As Kellogg and Bell (1904) have shown, the variation is continuous from large pronotal dashes through small ones to none at all. The presence or absence of pronotal dashes serves to distinguish specimens in collections from some localities, but not in the San Francisco Bay region. Fin. 14.— Variation of pronotal pattern, a. i>, d, e, from pilley, Oregon,- c, Fail-field, Washington; /to j, Berkeley, California. TAISLE 4. — Elytral pattern in Hippodamia convergens. Locality. No. Spot- less. Some Spots absent. 12 spots pres- ent ; no con- tinence 44-5 H+. Var. e.\- ten- sa. ™* Var. qllill- cim- mac- nlata. i ilac- ialis pal- tern. 1+2 5+6 ^ Palo Alto, Cal. (K. and B) 1,033 P.ct, \ — P. ft. 6.09 P.ct. 93.6 P.rt. 1— P. ft. 0 P.i-t. 0 P.ct. 0 P.ct. 0 P.ct. 0 P.ct. 1— P.ct. 1— 'i— Kamiack Butte, Wash 15 415 1— 1 5 :{ 5.91 6.3 30.2 3.7 s.o 1— 1— 0 0 0 Marsh Hill, Fair- field Wash 1,406 1— 8.6 '.9 2.4 9.8 0 0 0 0 0 TABLE 5. — Elytral pattern in Hippodamia convergens. Locality. No. Per cent spotless. Per cent some spots ab- sent ; no con- fluence. Per cent typical spots present; no con- fluence. Percent some spots con- tinent, Chewelah, Wash 191 0 27.2 1— 72.2 Alount Carleton ^'ash . 359 0 1— 0 99+ S pokane and Cheney, Wash. . Fairfleld Wash 107 1 ,406 0 1— 5S . S 33.7 5.6 8.6 :!5 5 50.2 sii-ptoc Buttt- Wash 0 81.3 0 is i; Kamiack Butte, Wash < lolclenclale \\'a^li. ... 15,415 159 1— 0 10.7 11.9 10.7 84.9 4S.4 :{.! Portland Ore"'. 10 0 10 90 0 Dillev ( )reg 895 94.9 l.i' 1— 0 Berkelev Cal 6T3 45. ti 4.0 48.8 1— Oakland, <_'al., A 632 9 2 1— 0 0 Oakland ( 'al B 76 ;;: 5 0 68.4 0 Watsouville <'al 68 51.4 0 48.5 0 Santa ( 'lara Valley, Cal Mendocino County ('al 425 84 2.3 0 6.3 3.5 90 :; 9(1.4 ] 1) Licking Fork Oal . . 17 5.8 70.5 23.5 0 lledlands, Cal San Diego Cal 71 4 12.0 75 1.4 0 S5.!l 25 0 0 I'alo \ltii Cal 1,033 1— (i.09 98. (i 1.7 San .lose Cal 267 1.1 1.8 97.0 0. Coolidfie. N. Mex Tepexpam, Mexico Bartlesville ( )kla 75 201 19 0 5 0 14. li 17 0 32.0 78 100 53. :i 1— 0 Ston\" I(ake !\Iich 20 0 0 95 0 Cold Spring Harbor, N.Y . . . 203 0 1— 99+ 1— 26 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. TABLE 6. — Variation of pronotiim of Hippodamia convergens. Locality. Elytra! spots. No. With dash. Marginal dash. Margin. With- out dash. Con- fluent cepha- lad. Con- fluent laterad Con- fluent cepha- ladand laterad Com- plete and not con- fluent With dash. 1 11CO111- plete ceph- alad. Incoin- lucom- plete plete ceph- laterad alad and lat- erad. Chewelah, \\':tsh Do Some a b- xent. Some eon- fluent. Do Present Some a b- sent. Some eon- fluent. Present Some a I) - sent. Some con- fluent. . ..Do 52 138 139 6 63 38 14 94 US 200 466 107 118 2,372 100 100 154 8 38 850 383 42 750 33 28 643 32 13 84 12 4 58 4 24 11 40 1!) 12 19 202 200 P. ct. 94 100 99 100 94 92 100 97 100 93 98 96 9' 96 97 99 100— 100 100 100- 100 98 42 100 (IS 100 97 100 100 100 100 100 50 100 73 0 37 25 100 100 UK) P.ct. U 0 1 0 (•> 8 0 3 0 7 2 4 • > 4 3 1 1— 0 0 1— 1— • > 58 0 32 0 3 0 0 0 0 0 50 0 27 100 63 75 0 0 0 P. ct. 6 0 1 0 3 3 0 1 1 1 2 2 1 4 0 0 0 12 5 :: 0 0 1— 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 P. ct. II 0 0 0 0 I) 0 0 (1 0 0 0 0 1— 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 P. ct. 0 0 0 0 0 0 0 0 0 0 0 0 0 1— 0 0 0 0 2 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 P. ct. 88 86 41 100 97 ti:! 100 98 11 89 89 86 88 P. ct. 0 u 0 0 0 5 0 1 0 i 0 0 0 P. ct. P. ct. 6 0 14 0 58 0 0 0 0 0 29 0 0 0 0 0 88 9 0 0 9 0 12 0 11 0 Mount ( 'arlton,Wash Spokane a nd Chenev,Wash . Do Uo Fail-field, Wash.. 1907. . l>o. 1 >o Fail-field, Wash.. 1H07.. Steptoe Hutte, \\'ash . . Some a b- sent. Some con- tinent. Present Some a b - sent. Spots land ti absent. Some con- fluent. Present Absent . Ho.. Kamiaek liutte, Wash Do . . . .... 96 1 )( > 99 94 100— 76 74 82 100 7<> HI 94 3 700 97 85 100 100 100 100 0 100 73 100 21 92 100 100 99 0 5 1— 0 0 0 0 0 9 0 7 0 0 0 0 0 0 0 0 18 0 21 8 0 0 0 1 0 1 0 o o 0 12 10 0 5 1 — 0 M 5 19 4 71 0 li 18 71 0 0 0 3 15 0 (1 0 o o o o 0 0 0 100 0 0 0 9 0 0 11 47 0 0 0 0 0 0 0 0 I ). > . . (ioldendale, Wash Portland, ( >reg. . . Oil lev, Oreg Some ab- sent. Absent Present Some a b - sent. Absent Present. . . . Do . Herkelev and Oakland, ('al. Do Do.. Watsonvillc, Cal .. Do Absent Santa Clara Vallev, Cal Do.. Present Some a b- sent. Absent ... . Present Some a b- sent. Present.. . . Do Do Mendocino ( 'onnty, ( 'al Licking Fork, Cal Do.. Kedlands, Cal. . 1 i, , Absent.. . . ( 'ooliilLie, N. Mex Do.. . . Present Some a b- sent. Some con- fluent. . . Do Do . 1 Hirango * 'olo Mill'ord, I'tali . . . Do Stonv Lake Mich . Present Cold Spi-insi' llarlior. N.Y. . Tepexpam, Mexico Total Do Do.. 7,824 89 11 1 1— 1— 80 2 8 8 1 i HIPPODAMIA. 27 Fig. 15 shows the variation in the pattern in the San Francisco Bay region. Subspecies of the mountains and high latitudes has a larger percentage of the characteristics of J -f 3, v + 3 +1, 4 + 5, and the absence of pronotal dashes and reduction of the light margin in the pronotum. I a Fi(i. 1">. — Variation in color-pattern ot head ol' Hippodamia runrrriji-nx. <• = mode: t'l.">,41.~> individuals I'roni the to)) 3 1 5 9 Kamiark Butte, Wash Palo Alto Cal 15,415 1 0:i:{ P. ct. 0 06 .29 P. Cl. P. <•!. O.'JO 1.08 78 1 4(> P. <•/. :$.fi4 .78 P. t-t. 8.62 87 /'. ,-f. •J7 -•; •'•' :; P. <•/. 59.2 93 9 Cold Spring Harbor, N. Y. . 202 0 0 0 0 0 1- 99 + 30 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. The variation of the number of spots is shown in table 7. At Palo Alto (data from Kellogg and Bell, 1904) and at Cold Spring Harbor the polygons are not regular, but show a normal condition of 6 spots, with a scattering few having aberrant numbers. I have grouped 25 specimens, said by Kellogg and Bell to have 1 or more extra spots on one or both sides, with those having 6 spots, because the irregular position of these spots and my experience with living specimens makes me believe that such spots are nearly always, except when at e and q, of ontogenetic origin. The poly- 17.— Relative frequency <>t the several spots in the semis /fi/ipodaniia. Broken line, col- lection of KellOKs and Bell from Palo Alto. California: entire lin« . collection from Kamiack Butte, Washington. gon for Kamiack Butte might have shown about 0.5 per cent, with a seventh spot e, but I have not put these in, because no accurate account can be made of spot e, as it may be entirely covered with pigment from spot 3 and the scutellar mark. Now, in contrast to the condition at Palo Alto and Cold Spring Harbor, we have at Kamiack Butte a polygon not greatly un- like the half-Galton curves found in some petal-counts. Yet the real nature of the variation in the three cases must be the same. Such an instance shows us that the distinction between continuous aud discontinuous varia- tion is not a sharp one. The order in which the spots disappear has proved to be a matter of great interest. I have tabulated the data given by Kellogg and Bell (1904) Palo Alto, California, in table 8, and fig. 17 (broken line), but have numbered the spots according to Weise. There is no very pronounced order of disappearance here, the most prominent feature being the persist- ence of spot 2. But at Kamiack Butte (table 8, fig. 17) , a very different condition prevails. Here spot 1 exhibits a strong tendency towards disap- pearance, and this happens in spite of the fact that it is in company with several varieties having spot 1 unusually large and in some cases confluent with spot 3. Spot 6 follows it in order of non-development and then spot 2, which at Palo Alto was most persistent. It is evident, then, that not only is there a definite order, but that this differs from place to place. The HIPPODAMIA. 31 spots are, therefore, to a considerable degree independently variable units. Yet if we take into consideration the combination of spots, we shall find there is some ' ' coupling. ' ' Spot 1 is most frequently the only spot absent, while spot 3 is most frequently absent as one of three. It is quite possible that the order of disappearance would be just the reverse of the order of persistence. Table 9 and fig. 18 are designed to test this matter and also the question of coupling between spots. Along the horizontal scale I have represented the various combinations of spots in the order of frequency of their disappearance. The solid line indicates the frequency of these sev- A Spots 11211011111311224134 6 22 34254 5363626 3 66 66 5 154112 4 5234 5 45 Lacking 3583664327313241241 168 1.57 73 72 41 32 20 25 19 18 14 10 9 8 7 7 6 5 4 1 Having 04310241 0109101 33 90 2 5 613 66 04 1 FIG. is.— Correlation of loss <>f spots. 2 3 5 r, i i 8 26 1 3 5 346 4 0 0 0 = 5073 1 4 6 = 211 eral patterns. Now, in the dotted line I have represented the frequency of the same combination of spots present. We see from it that spot 4 alone is less frequently lacking than spot 2 alone, yet spot 2 alone more frequently persists. Similarly, spots 5-6 are less frequently lacking than 4-5, yet 4 and 5 each persist more frequently. The couplings of pairs of spots in disappearance in the order of frequency are 1 and 6, 1 and 2, 1 and 3. These pairs do not frequently persist alone, and the order of pairs in persistence is 4 and 5, 2 and 3, 3 and 5 ; these pairs are not frequently lacking. The coupling of trios in order of disappearance is 1-2-3, 1-4-6, 1-2-6 ; these groups rarely persist alone. Thus it is roughly true that the more persistent groups are less likely to be frequently disappearing groups. If the confluence of elytral spots 4 and 6 represented merely an overflow of pigment, because of increase of size of spots, it should be found in H. quinquesignata or boivditchi, yet it is only rarely met with in cases of enlarged 4 -f 5 in the mountains. It is more frequently found in Califor- nia in regions where the spotless variety is present. 32 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. The confluence of spots 5 and 6 is an independent variation. The speci- men that Major Casey has from Sonoma County, California, and the one I have from Stony Lake, Michigan, are well marked. It is very rare, with- out intergrading conditions, and constitutes a distinct position of organic stability. TABLE 8. — Spots lacking in Hippodamia convergens. [Percentage of hectics in which each spot is lacking.] 1 Locality. No. Spot 1. Spot 2. Spot 3. Spot 4. Spot 5. Spot 6. Palo \lto ( 1jil .. ... 1,033 :{.5 0.9 2 7 :', 5 :; 1 :; !i Kainiaek Buttc. Wash 15,415 Cold .Spring Harbor, X. Y 202 'IV| ic \pain Mexico . 200 :!ti.2 7.7 0 0 110 7.0 4.1 0 7.0 2.4 0 12.0 1.5 0 13.0 9.3 j 19 0 The loss of the pronotal dashes (fig. 14) and the extension of the black pigment on the pronotum is well marked in specimens from California Strangely enough, they affect the spotless beetles most. One might conclude that the result is compensatory for the lack of pigment in the elytra, but in Oregon, where spotted elytra are found, the black pattern of the pronotum is reduced, as shown in fig. 14, a to e. There are prob- ably, therefore, two different causes for the reduction of pigment in the pronota and in the elytra. The only case of an opening of the pronotal dashes (fig. 14) to the margin laterad without communication cephalad was at Fairfield, where the forward opening is not nearly as common as in Oregon. Apparently the line of development is different in eastern Wash- ington from that of western Oregon. TABLE 9. — Combinations of spots present and absent in H. convergens from Fairfield, Wash. Total. Present. Spots. . \liscnt. Total. | P. el. of sum of totals. In sixes. In fives. In fours In trios. In pairs. Sin- giy. Sin- In -ly. pairs. In trios. In In fours, rives. In sixes. 9,828 11,779 14,233 15,191 15,048 13,992 Av 9,121 9,121 9,121 ! 1.121 9,121 9,121 615 4,166 3,871 4,19] 4,190 3,957 69 1.131 1175 1,247 1,265 621 13 283 219 512 387 269 ID 69 44 107 80 22 0 9 • i 13 5 •> 1 9 5 1 (i :i,5s:; l,_'5s 32 196 327 .T>2 7 80 S 62 211 70li 54S 27X 342 49 174 292 15(i 32 97 23 122 29 5!) Ill Xli 27 144 30 10 10 10 10 10 1(1 5,587 636 1,182 224 367 1,423 36.2 (.1 7.7 1.5 2.4 9.3 9,121 3,49K.3 884.7 280.5 55.3 5.3 6' r.1.7 442.3 2X0.5 110.7 26.7 10 9,419 JO. 2 CORRELATION. The correlation between the confluence of spots 4 and 5 with the con- fluence of spot 3 and the mark J- is represented in table 10, and of 4 and 5 with the confluence of 1 and 3 in table 11. These tables show that the cor- relation falls far short of the current notion, expressed in the systematic literature, that these confluences are definitely coupled. An examination of table 6 shows that where black spots on the elytra are absent the pronotal dashes are also absent in a large number of cases HIPPODAMIA. 33 in some California localities, but not in others. At Coolidge, New Mexico, where spots are confluent, the pronotal dashes are absent in a large propor- tion of cases and are present where there is no confluence. This is also noticeable in some small lots from Colorado and Utah, but in the Pacific States this difference is not marked. TABLE 10. — Correlation of relation of spots % and 3 with relation of spots 4 and 5 in Hippodamia convergens at Fairfield, Washington. 4 ami ">. Relation, '._, anil 3. Con flu- fiit. Transi- tion. Sepa- rate. Absent. Ti >tal. Confluent * 177 2 22 1 202 Transition . ... 1 I) 6 1 8 Separate 1 1 1 t'J9 4 118 Absent 0 0 5 tl 6 Total... 1112 3 132 7 334 * Quindecimsiffnata and Var. oltxo/rla. TABLE 11. — Correlation of relation of spots 1 and 3 with relation of spots 4 and r> hi Hippodamia convergens at Fairfield, Washington. 4 and 5. Relation, 1 and 3. Fused. Transi- tion. Sepa- rate. Abselil. Total. Fused *64 0 0 0 64 Transition 1 0 0 0 1 Separate f!20 3 159 0 1S2 Absent 7 0 74 g6 87 Total 192 3 133 (1 334 Var. quinquesignata. t Var. cnseyi. t Convergens. \ Var. obsolete. The lateral extension of the black area on the pronotum to the margin, so as to make ''margin incomplete laterad," is associated, strangely enough, with both spotless elytra and elytra with spots confluent. We shall see later that this characteristic is subject to modification, so that the great variation between the several localities in eastern Washington which differ in altitude is readily understood. In fact, a lot of hibernating beetles from Fairfield in 1907 showed far fewer specimens with the margin incomplete laterad than those of another year. The cephalic extension of the black area of the pronotum causing an incomplete margin cephalad is only common in California and New Mexico with beetles lacking the pronotal dashes, but that it is not a necessary con- sequence of the absence of the dash is shown by the specimens from Mil- ford, Utah. We have, then, in different localities, a difference in the correlation and the order of development of the pronotum and the elytra. 34 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. MEASUREMENTS. I have measured, by means of the eye-piece micrometer, the degree of separateness or confluence of the several spots in a lot of beetles from Fair- filed, Washington. Fig. 19 gives the result in the case of spots 4 and 5. The micrometer units of 0.625 mm. have been used without change to avoid the introduction of errors of reduction. In order to avoid errors of com- binations the original classes are used, even though it makes the polygons somewhat irregular. The reader may make his own combination of classes mentally. Of course, since it was necessary to measure a projection of the curved surface and as the distance measured was not always at the same focal distance, there are errors involved. Experimental testing, however, shows that the errors are too slight to affect the significant features of the (n) The distance between and confluence of ppots 4 and 5 (in units of .625 millimeters.) (6) Males, (c) Females, (f?) All individuals. (a) 14 13 12 11 10 '.18765432101 23 1 5 C 7 S 9 10 11 12 13 14 15 16 (b) 0 1 0 2 2 6 10 11 S 84 Ox! 20 1 0 1 3 4 811 15 14 13 13 2321 0 = 1.30 (c)000j4696 13 11 S 7 61210345 10 ]<; 21 13 7 7 (i 1 0 1 0 = 169 (rf) 0 1 0 3 6 13 19 17 22 19 13 7 S 3 2 2 0 5 7 9 IS :>,\ 3C, 27 20 20 8 4 2 2 0 = 324 FIG. 19. — Variation ol the distance between and confluence of spots 4 and 5 in Hippodamia convergcns at Fuirnekl, Washington. polygons. A novel method has been used in the construction of the curves. 0 is the case where the spots are separated by a distance less than 0.5 unit or confluent to the extent of having a connecting strand less than 0.5 unit in thickness. Increase of distance between the two spots is plotted to the left. Increase in breadth of the connecting strand pigment is plotted to the right. It will be noticed that the sexes do not materially differ ; that the dis- tance apart is more variable than the thickness of the connecting band; that the polygon is bimodal; and that the transitional cases fall for the most part into one or the other of the two polygons. HIPPODAMIA. 35 Fig. 20 shows the variation in the relation of the mark ± to spot 3. The same conditions hold as before. ( ;unl spot 3. (6) Mules. (<•) Females. (rf) All individuals. 17 in i.-, (ft) 005 (»•) 0 1 1 (fl) 0 1 G 14 13 12 11 10 987654321 0123456789 10 II 12131415161718192031 3 03313661 10 5248300136147 1314 !>:• 11 7 9 0 2 o o 0 0 0 = 152 2328505565438532022459 13 8 10 10 15 986100010 = 170 5 3 5 5 G 9 11 11 7 16 9 5 12 9 6 2 0 3 5 10 7 13202124272610 li (.; 3 0 0 0 1 0 = 327 FIG. 20. — Variation of the distance between and confluence of the mark % and spot 3 in Hippoclamia convcrgens at Fairfleld, Washington. Fig. 21 (of the relation of spots 1 and 3) gives a curve which is prob- ably trimodal — a mode each for the typical species, the variety caseyi (widely separate spots) and the variety quinquesignata (confluent spots). (a) (b) (<•) (d) The distance between and confluence of spots 1 and 2. Males. Females. All individuals. \ (a) 23 22 (ft) 00 (c) 01 («f) 0 1 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 1 2 3 4 5 6 7 8 9 10 11 12 0131362203224 11 21 16 16 21 3 0100011115014 0=111 0114V 42427404 15 20 17 12 2120001120742101 0=120 0 1 4 o 10 10 4 G 2 10 6 2 8 26 41 3329 4 2 5 0 1 0 1 1 3 1 S 5 7 1 15 0 = 213 Fi«. 21— Yiuu.tio.1 ut tlic distance between and confluence of the spots 1 and 3 in Hippodamia convergens at Fairfleld, Washington. The reality of the varieties is conclusively shown in fig. 21. They are not distinguished by sharply limited unit-characters, but are centers of variation. 36 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. The variability in diameter of spot 1, shown in fig. 22, is a partial expla- nation of the previous polygon. Here the variety caseyi and variety quin- quesignata include these individuals at or near 15 units, while the typical specimens are at or near 5 units. ("} 0 ('') 0 ('•) i> (<() 0 («) Diameter of spot 1. (6) Males. (c) Females. (rf) AlLinriividual>. 12 •-! -1 3 5 5 0 3456789 10 11 12 13 14 15 1G 17 IN l.i :J'i :.M ;.'•.' 3 (i •"! 311122 5 12 19 22 15 10 2120 0— IF, 3 37331445 15 10 13 III 14 0 4 0 0 1 0=129 G 0 10 G 4 55 6 7 20 31 32 42 29 10 6 1 ~ 1 0=24r, FK;. 22.— Variation in tin- diameter of spot 1 in /[iiijiui/itiitia <-1MS 0 54.03 ± .20 13 50.71 ± 0.22 ">l.77 + .23 Var. quinquesignata ,'>4.06 -t- 39 52,41 ± .38 HIPPODAMIA. 37 MODIFICATION. Subjection of the prepupa and pupa to an ordinary poultry incubator at 40° C. resulted in no appreciable modification, nor did increase or decrease in humidity. But an increase of pigment was obtained by expos- ing the prepupa and pupa to the cold of a refrigerator (5° to 15° C.), a cellar (15° to 17° C.), and the intermittent temperature of an ordinary room where the temperature dropped during the winter months from 12° to 18° C. by day to 5° C. at night. The greatest degree of pigmentation resulted in the last case. The result from one experiment, in which the prepupa3 of typical Hippodamia convergens were subjected to a freezing tempera- <«> Diameter of spot 3. / ture out of doors for two nights, ((^ ^envu^ was the elytral pattern shown in </'"- show a correspondence to some of the varieties in nature ; other features are produced only by artificial modification, especially the pigment line between 1 and 4, which is not the same as the vitta in Hippodamia spuria var. or H. apicalis var. HEREDITY. Every variety or noticeable variation which has been tested is inher- itable in some degree. Even though the variety has been brought from a different region, it has maintained itself in our vivarium. The heredity of spot 1+3 shows a nearly perfect segregation. There is a lower degree of segregation in the confluence of 4 and 5. The closest approach to blended inheritance is in the size of the pronotal dash and the incomplete lateral margin of the pronotum. Yet in each of these cases it 1 2 3 4 5 3 3 3 3 G G 7 8 9 10 11 12 13 G 11 8 7 7 4 1 0 = 55 5 8 17 13 5 3 0 0 = 03 5 11 19 25 20 12 7 1 0 = 113 38 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. is clear that the end positions, that of no spot verms spot of normal size in the one case, and lateral margin wholly incomplete versus normal in the other case, are more stable conditions; otherwise we should probably have, as a modal condition, a somewhat shorter pronotal dash and a narrower margin, instead of the persistence of a variety in company with a parent species. , FIG. 25.— Offspring of 29.3, , exposed to inter- mittent cold when a prepupa. FIG. 26. — A new pattern, produced by inter- mittent exposure to cold, which is not found in nature. From Hippodamia conreryens. () 0 1 2 4 12 15 37' 35 20881000001 0 = 139 (f) 0 0 0 7 12 28 35 44 20 3 1 0 0 0 0 0 0 0 0 = 150 (eO 0 1 2 11 24 43 72 7'J 40 11 41000001 0=M-J FIG. 24. — Variation of distance from the suture through spot 3 in Hippndti- mia converyens at Fairfleld, Washington. (a) Distance from suture through spot 3. (b) Males. (<•) Females. (d) All individuals. Owing to the small number of extended pedigrees, the extent of domi- nance and segregation in these beetles mast be examined in large part by the comparison of the numbers in fraternities, only one parent or two parents and one grandparent of which are known, with the theoretical expectations under the several assumptions. In table 13 I have collected, for convenience, the proportions to be expected under the various condi- tions and assumptions. Thus, if a characteristic gives mixed broods when interbred, it is not recessive, even if the allelomorph is more strongly inher- ited. It is principally upon this criterion that so many of the pedigrees fail to be simply Mendelian. Where all the spots are absent (table 15), in the parent, we find this condition ordinarily in the offspring. In the progeny of the 12 females of this kind which had mated in nature, we find that not infrequently a few HIPPODAMIA. 39 of the progeny have a few small spots. None were fully spotted. Com- bining the progeny of these 12 (81 in all) and weighting small, hazy transi- tional spots as one-half, we find spots 1 present 4, spot 6 present 5 + I, spot 4 present 9 + |, spot 3 present 8 -|- £, spot 2 present 11 + f , spot 5 present 16 + f . This higher degree of persistence of spots 2 and 5 and the lower degree of 1 and 6 we have also seen in the variation results in fig. 18. Unfortunately I have no cross of typically spotted and spotless. In 149 we have spotless x; spots 3 and 5. The progeny showed these spots in 6 + iand 6-j-|, respectively, in the 11 offspring. A spotless individual never gave fully-spotted offspring, nor did fully-spotted indi- viduals ever give spotless offspring. Partly-spotted patterns can not be considered a heterozygous condition, because in one experiment where the parents (387) were 2 partly spotted individuals, the 20 offspring were spotless or partly so. The spots which were present most fre- quently in the offspring were present in the mother, but the particular degree of spottedness was not stable, for several were wholly spotless and several were provided with more and better spots than the parents. We may conclude that in spottedness : ( spotlessness simple Mendelian heredity does not prevail, but that crossing gives an intermediate con- dition which, while not acting like a heterozygote, is yet unstable and contributes to both conditions. TABLE 13. —Percentage of progeny to inherit a characteristic on simple Mendelian expectation. Characteristic is — Father. Dominant. Trimorphic. Recessive. Mothei A A. Mother A <> Mother A A or A b. Mother ' < . Mother a a. A A 100 10(1 100 A asal mark very prominent in M, normal In progeny. (1) In M 4 is verv near to .">. ( )f :! others subjected to cold 1 had I tr. ;>. (5) In M' 1 + 3 became slightly confluent alter several days. 42 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. TABLE 15. — Heredity in Hippodamia convergens. ( See bottom of p. 41.) Father. Mother. / Nor- mal. J and 3 1 and 3. 4 and 5. 4 and f>. Ex- tensa. Spot 1. No. Offspring of— No. Off- spring of— Tran- Tran- sition, sition. + Tran- sition. Tran- sition. + Tran- sition. Ab- sent. It R K C S H 5183 5294 430' + CS H O P A .-•7 ' + 0 5x • + O I1 A r A D D 1> D I) ._,_,,; B CSH C S H CSH F F 41'" 411J CSH F 112- 406! 1117' in;;- 4GS- F 435^ CSH F D 201" 259' P 413f 417' 419' 421 1, 357 i + F 408' 41S' F 420t 427 - F 390' 4.VJ ' 391 ' F H92 ' 393J 386J 387t :;::;• + F 4293 IH' + C ny F 4093 1153 F F F F F F 422.' + F •157' 4643 3823 439.J R R R CS 11 <) H OB 357 + C S H i ) PA 1' A + O P A + 0 PA P A D 1) D 1) D 173 B CS H CS 11 CS H F F 356 356 CSH F 369 364 369 369 403 F 375 CSH F D 151 151 F 352 352 352 352 11)3 + F 357 F F 360 32 352 352 352 F 357 357 F 352 352 F 354 390 354 F 351 351 351 351 F 373 C 114 F 366 366 F F F F F F 390 422 F 382 4 13 5 6 6 3 39 3 1 8 1 2 2 2 8 25 7 17 3 5 3 1 10 7 25 9 14 11 12 7 22 10 12 3 19 12 24 £i 25 7 9 13 13 2 1 5 21 10 20 5 7 16 17 15 3 5 39 11 26 14 2 14 4 11 12 7 2 1 1 6 5 3 1 16 21 20 7 13 M 3 M 13 M 5 M 6 F 2 FM3 M39 1 f M3 M 1 fM 8 f M 1 M2 M2 M 2 MS M 25 M6 M 17 F M 3 Mo 1 M M 1 10 •> 3 1 3 4 11 .... 1 3 M5 20 F M 8 KM 14 2 F M M2 1 FM7 11 1 2 M 1 3 i F 1 FM10 FM2 FM4 F,M3 •• 3 2 6 Fl 3 M 4 4 8 M5 3 M 13 FM10 1 M 1 1 3 1 1 4 1 4 .> 1 1 M 4 10 .... •• 1 M (i I M '2 " 3 9 FM4 M !» i r M12 2 M13 FM13 F M|2 FM1 F M5 2 •) 1 FM13 FM2 FM 1 FM5 8 .... M"-2 M .... M 5 '2" i M7 i f 11 K M S M 19 i M 5 M 5 FMG FM15 M17 FM15 FM3 F M HI KM 10 FM1S 5 Fl •> M FM7 FMlli MIC FM15 FM3 FM5 10 M3 M 7 M M10 FM15 F M 3 ... F M 5 1 4 F 2 4 F3 M 6 1 F*4 K M 2 1 2 M 14 'M 7 FM18 1 M 5 1 4 3 7 1 3 3 1 FM11 FM 3 3 M F 9 FMd 4 F M 9 'i M3 5 M4 F M 1 M Fl M :; F .M 10 2 6 1 M 1 2 Si .... FM6 3 F 1 6 FM7 1 M M 1 M3 M5 Ml M 1 FM14 FM21 FM20 FM7 FM13 M 1 M 1 3 M5 3 M .... 1 M 4 M 2 1 M 1 .... Ml Ml Ml FM15 FM21 FM17 M7 FM13 1 'F" M21 F20 2 F M 1 "2" F F HIPPODAMIA. TABLE 15.— Heredity in Hippodamia convergens — Continued. 43 Spot -2. Spot 3. Spot 4. Spot 5. Spot (i. Pronotul dash. I'ronotal margin incomplete. Remarks. Tran- sition. Ab- sent. Tran- sition. Ab- sent. Tran- sition Ab- sent. Tran- sition Ab- sent. Tran- sition Ab- sent. Small Ab- sent. Tran- sition open Ceph- alad. Lat- eral 1. M * M * (1) f* M* M * f* M * I'ronoluin open: laterad also in m. M* M * (-') Spot 6 in ott'sprins? decidedly small. M t (3) M * 1 offspring sk F* F* (4) 2 offspring ^ F* F* M * (5) 'A F missing. It was "almost spotless." (5) F* M * « 3 offspring >(c (3) F*M * F* M * f * M * ] absent on left side 1 offspring ^ M * F * M * 1 offspring >(c 1 offspring *. On K. 1 a p.; on L. 1 tr. F* M 2 1 2 M4 . 4 1 M M 3 Ml r M s I'M M •_> M '2 M 1 M K M13 MG Ml.") FM2 M5 .... 5 1 M M3 M 1 f M S f M M2 M2 M M 8 M10 M 6 M 15 Fl M 5 if' M 3 1 M M3 Ml r M s i M2 M 2 M2 MS M15 M6 Mlo F2 M 5 1 1 1 .... 1 1 M3 M 1 f M S r M i M2 M 1 Ml M7 M17 M6 M12 F2 M4 M3 M 1 fMS f Ml M2 M2 M 'I 1 1 M ! 2 1 M 8 f M 1 2 2 1 1 5 fl M t 1 M M S f] M 2 M 1 "i' 1 MS M 19 M6 M 17 F M 3 M5 3 M 5 F 1 M4 . . . 1 4 ."> 5 M M •> 2 2 1 1 1 1 1 M "i 1 M 'i M 1 Y •_> 2 Ml 2 1 M M 1 I 4 •> 3 F M 1 M9 FM10 1 1 M 1 6 1 4 M 1 5 1 3 1 M 1 1 1 3 1 <> 0 1 3 1 Fl 1 1 M M M5 F;MG 1 1 3 M 1 M Mil M 4 1 M FM6 M 1 F F M 4 M 2 F F M 2 5 1 3 Fl M2 f Mil M 7 17 3 2 M~3 fl FM6 12 2 4 13 4 M17 3 5 M 4 i 3 7 1 f3 4 4 1 M 17 M 1 4 4 F M 4 M9 Ml Mo FM7 M5 S M 12 M 2 F M 2 "3' 3 M 15 F M3 M2 M 7 3 8 2 j M 3 F M 12 9 6 3 11 4 9 1 1 .... M3 •> ' i M M 3 FM7 M 2 1 2 1 1 FMil 2 2 F3 1 1 2 F M~10 f 1 5 M 3 M M f> M7 3 "2' Irs i 1 F4 F 1 1 M3 M42 M 1 .") 17 10 F M 5 M13 . M i 1 1 4 M 1 Ml FM 12 18 FM16 FM2 M9 i' M F .... 1 44 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. The convergent dashes upon the pronotum were shown by Kellogg and Bell (1904) to vary in size down to absence, and my observations confirm this. No high degree of segregation is expected, then, in inheritance, and none is found. Where the parents have intermediate dashes, the progeny also have, although some may be just large enough not to be classed as small in the table. The inheritance, then, is blending, but with so high a variability that both absence and a considerable size may occur in one fraternity. The condition of pronotal dashes communicating with the light margin is inherited in an average of 27 per cent, when the one only known parent has the characteristic. Where the other parent does not have it, as with 149?, none of the offspring have such a communication. Where both parents have the communication, as in 435, then 6.5 in 11 possessed it. There is, therefore, neither dominance nor trimorphic heredity, but sub- ponderance. The data upon the incomplete margin cephalad is only adequate to show that this character is inheritable and shows some segregation. The incomplete margin laterad is classified upon an unsatisfactory basis, since so many individuals have the margin almost incomplete. We have further- more to deal with some modification. Yet it is clear that the character- istic is inherited in part and it seems toigive us continuous variation. It is quite possible that intermediate degrees are discriminated against in devel- opment and that either a large lateral projection or a very slight one is favored, as the facts of variation and distribution would cause us to suspect. If this species as here constituted is ever to be dismembered, it is most probable (aside from the removal of variety moesta) that variety quinque- signata and variety caseyi will be removed as H. quinquesignata. There is some ground for this step, although, considering everything, I have not chosen to make it. My experiments show a partial intersterility between specimens of variety caseyi of full pigmentation and eastern specimens of H.convcryens. But there was interfertility between specimens with the same confluence in a lesser degree and having spot 1 small (the variety dejecta] and eastern specimens of H. convergens. I have only hesitated from the division because I do not know whether variety caseyi and variety defecta may not be interfertile. The presence of intergrades causes me to suspect that this is not only so, but that they habitually interbreed. At any rate, this seems to be a point at which this species may in the future divide to give us two species, even though it may not yet have done so. HIPPODAMIA. 45 3036 a. Hippodamia bowditchi, new species. Distribution: Northern Rocky Mountains. The type (fig. 28, rf) and 5 cotypes (fig. 28, 6, c, c,f, g) of this species were collected from goldenrod August 12, 1904, on the bank of the river at St. Maries, Idaho. The pattern is comparable to that of H. convergens and vari- ety quinquesignata and would be considered a variety of that species but for the distinct hiatus between the 4 specimens of variety quinquesignata and 11 of other varieties of H. convergens, taken in the same lot, on the one hand, and the 6 in question on the other hand. The black bands of quin- quesignata are here much enlarged; the intermediate band is thicker in the baso-apical direction than the space between it and the basal band. The wide basal band is nearly uniform in width and thus shows little evidence of its component spots. It shows no rounded projection at the position of spot 3. Spot 2 is absent and the apical spot is much enlarged. The inter- mediate band is so enlarged as to obscure its compound nature and its apical margin shows no angulation. Its basal margin has an angulation, but no restriction is produced in the band. The band extends far toward both the suture and the costal margin and shows a straightened edge in these directions. 9 . 2s.— Variation of color-pattern of Hijipodamin boirditchi. d= mode; 6 to . Var. y. Spots 1, 4 + 5 + 6. Rare, Idaho. Nowhere established (fig. 29, n). Var. e. Spots 1 absent, 4, 5, and 6 reduced. 2 in 759 at Fairfield, Washington; Colorado (fig. 29, g). Subspecies of Colorado: Reduction is most manifest in Colorado, a par- allel to the condition in Hippodamia apicalis in Colorado. Whether it is local to some part of Colorado, as is probable, or not, can not be told because of the former pernicious habit of labeling by States only. The confluence manifests itself in each individual either as 1+4 +5 or 5+4+6, and never at Fairfield involving all 4 spots even in a slight degree. The two varieties are therefore distinct in this locality. In Vancouver Island, however, the combination is found. HIPPODAMIA. 47 The pronotum is much less frequently subject than that of H. conver- gens (see fig. 30) to a reduction of pigment in such a way that the con- vergent light mark is confluent with the light margin. This was found in FIG. 29.— Variation of elytral pattern in Hippor/amiaspuria. a FIG. 30.— Variation of pronotal pattern in Hippodfimia xpuria. Mode = c tx> .?. FIG. 31.— Variation of head-pattern of Hippodfimia xpuria from Dilley, Oregon. 6 = mode. only 5 in 759 at Fairfield. On the other hand, there was one specimen in the Fairfield lot in which the convergent mark extended to the caudal margin of the pronotum and another in which it very nearly did so. In each of 48 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. these cases the pronotum was lacking pigment at an independent center— the caudal margin of the pronotum. The variation of the color-pattern of the head is given in fig. 31- HEREDITY. Table 16 epitomizes the results of the heredity experiments with respect to the elytral pattern. No. 368 ? has a pronotum with pronotal dashes open to the margin cephalad. None of the 18 offspring possess the character- istic. Since the same characteristic is inherited in H. convert/ens (375?), it is probable that its lack in the offspring here is the result of recessive- ness or subponderance. TABLE 1(>. — Heredity in Hippodamia spuria. [For description of form of table and abbreviations sec table 15, pp. 42, 13.] Father. Mother. No. Sutnral pigment. 1 and 5. 1 and "i. ."> and fi. No. oil- spring , No. of— ( .11- sprintr of-- Short. Jntcr- ine- diale. Long Sepa- rate. Ap- proach Tran- sition. + Sepa. arate. Tran- si- tion. '"I" Sepax rate. Tran- sition. • F 1) 2i!2 ' UK) • + ? :-;s:; < + ? 4iil ' 4 S'.l • :;so < 4373 112 ' 44,5 !. |:!2 ' 433^ F 3«S , D 227 ; 227 2H2 ; F 400 } F 383 ; 3S3 4 M .') FM1 1 1 1 4 FM 5 FM2 F M 5 FM4 1 3 1 K:; In the elytral pattern the table shows that the sutural mark is inherited segregately, with some intergradation. We have two centers of variation, but not two unit-characters. The long sutural mark is neither recessive nor dominant, nor is there any regular dominance in the relation of spots 4 and 5. There seems to be a preponderance of short sutural pigment and of 4 + 5, which are more common than the opposed conditions. 3048. Hippodamia oregonensis Crotch. Distribution: Cascade Mountains. Although the pattern of this species (see fig. 32) is parallel to that of H. convergens var. quinquesignata, a collection of the 2 from Pinnacle Peak, Washington, shows a difference in form of body and shape of the spots. This makes it probable that it has been derived from H. spuria by the addition of the banding character. The possible influence of HIPPODAMIA. 49 the mountain environment will later be discussed in this connection. The vittate tendency is much more marked here, as might be expected, than in H. convergens var. quinquesignata, nearly all specimens showing some evidence of it. The absence of pronotal spots is another result to be expected from the mountain environment. The basal band is the chief differential from H. spuria. Specimens of that species differing only by absence of pronotal dashes are often labeled H. oregonensis. Such specimens should be considered a variety of H. spuria, as presence or absence of pronotal spots alone is a character of less distinc- tion in Hippodamia than presence or absence of the basal band. Form a. Spots £+1, 6+4+5, Pinnacle Peak, Washington. FIG. 32.— Variation in color-pattern in H. oregonensis. c = mode; 6, from Oregon; c to e, from Pinnacle Peak, Washington. FIG. 33.— Variation in color-pattern of H.cockerclli. a, from Colorado; 6, from Yellowstone Park. 3048 a. Hippo Jamia cockerelli, new species. Distribution: Colorado to Wyoming. This species resembles H. spuria, except in color-pattern (see fig. 33). This is remarkable in possessing vittaB and banding in combination. It is the Rocky Mountain equivalent of H. oregonensis of the Cascade Moun- tains. The two mountain systems have given two different mountain species derived from H. spuria. If intergrades are ever found it will prob- ably be at the convergence of the two mountain systems in British Columbia. The type is in the U. S. National Museum. It was collected by Prof. T. D. A. Cockerell, in honor of whom the species is named, in Cottonwood Gulch, on the west slope of the Sangre de Christo Range, in the Hudsonian Zone, in Saguache County, Colorado, August 4, 1887. There are also two specimens from Yellowstone Park in the collection of Mr. F. P. Bowditch. The three show very little variation. I have never seen any tendency to the formation of a basal band in H. spuria, although I have seen many specimens in the Western States. This species has apparently arisen discontinuously. The complete lack of intergrades I have taken to indicate that it is intersterile with H. spuria, since complete dominance, even if there be complete segregation, is not to be expected in coccinellid color-patterns. 50 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. 3046 b. Hippodamia dispar Casey. Distribution: Colorado. Of this very rare species I have seen only the type in Major Casey's col- lection, the specimen in the U. S. National Museum, and the one in the Leng collection (fig. 34). The 3 specimens show some variation in the rotundity of 4 + 5, but the combined spot is in all cases much rounded. The rounded enlargement of the 44 5 area in H. cockerelli points strongly to a derivation of H. dispar from H. cockerelli. A large series might reveal an amount of intergradation which would bring them into one species. But at present the few specimens known are distinct. FIG. ;>•!.— Color-pattern of II. dispar from Colorado. FIG. 3").— Variation of cl.vtral pattern in H. simiata. b — moilr; It to c, from ALumocla, California; a, variety crotclii. 6049. Hippodamia sinuata Mulsant. Mulsant's description seems to apply to the same beetles later named H. trivittata by Major Casey, and his type came from the same locality. This Californian species (see fig. 35), with the vittse complete, is, however, distinct from the other beetles often called H. sinuata, which have the vittae very strongly bowed in the apical half of the elytra, or incomplete or lacking; these latter are properly placed in H. spuria. Mr. Nunenmacher informs me that H. sinuata is restricted to the salt- marshes, and in my collection in California I failed to find them outside of the marshes. It is this fact, together with the lack of intergrades, that causes me to consider them specifically distinct. This isolation by its decrease of interbreeding would have afforded the opportunity to establish intersterility. Var. crotchi. The suture black for only one-third of the length. One specimen from San Fran- cisco in the collection of Mr. Leng. 3050. Hippodamia tredecimpunctata Linnaeus. Distribution: Europe, United States, and Canada. This species is the least variable of all the widely ranging species of Hippodamia. Yet large series show variation (see figs. 36 to 38). This is parallel in part to that given for Europe by Weise (1899) and Mulsant (1846). The names and letters employed by them will be used. While HIPPODAMIA. 51 some of these varieties as noted have not yet been found in America, most of them are to be expected. Var. padana Weise. Black pattern on pronotum small, less than twice the width of the light margin (cf. fig. 37). Elytral spots small. Var. signata Faldermann. Black pronotal pattern narrow. Elytra spotless. One from Washington, but with mark £. FIG. 36. — Variation in elytra! pattern in If. 6 = mode; a, from Washington; b, c, from Montana; (/, from California; c, from Massachusetts; f, from iculix a,l>, from Colorado. Variation of Color-pattern of head of //. par- enthesis. c = mode; c, (I, from Stony Lake, Michigan; e, from Toronto, Ontario. FIG. 42.— Variation of pronotal pattern ol If.partnthesis. c — mode; a to <; trom Dilley, Oregon; /to./, from atony Lake, Michigan. 54 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. 3051 a. Hippodamia apicalis Casey. Distribution: Colorado to California, Arizona to Washington. This species is closely allied to the the previous one (see fig. 43). It is given specific rank because of its smaller size, darker pronotum, different head and elytral pattern, and lack of intergrades. Since its range is within that of H. parenthesis, which it does not displace, it can not be a subspecies of it. In the following descriptions, a is a black mark along the suture at and near the apex. Typical pattern: Spots J+3, 1+2, 4+5+6+ a. (Fig. 43, e.) Var. Spots J+3, 1+2, 6+4+5+6+ o inclosing alight spot. Arizona. (Fig. 43, h.) Var. Spots J+3, 1+2+4+5+6+ a off. arctica Schneider. Nevada and Utah. (Fig. 43, flr.) Var. Spots J+3, 1, 5, 4+6+a. Idaho. (Fig. 43, i.) Var. Spots J+3, 1, 4, 5, 6+a. Colorado and Wyoming. (Fig. 43, c.) Var. Spots J+3, 1, 4, 6+a. Colorado and Wyoming. Var. Spots J+3, 5, 4+6, «. Colorado and Wyoming. Var. Spots J+3, 1, 4, 6, a. Colorado and Wyoming. Var. Spots J+3, 1, 4, a. Colorado and Wyoming. Var. Spots J+3, 1, a. Colorado and Wyoming. The pronotum is always much darker than in H. parenthesis, the caudal white spot being especially reduced. The head has 2 spots near the cephalic border in- stead-of 2 pyramids with their bases upon the border, as in H. parenthesis* FIQ, 43.— Variation of elytral pattern in H. icalix. e =- mode; a, from California; b, il, from Colorado; c from Wyoming; r,f, (/, from Nevada: h. from Arizona; /, from Idaho. Subspecies of Nevada and Utah: Spots never found missing and the vit- tate variety frequent. Subspecies of Colorado and Wyoming: The spots much reduced and some frequently lacking. Subspecies of Arizona: The black pattern much enlarged but not vittate. A further development in this direction is seen in Southern California in H. lengi. Hippodamia apicalis is more closely allied to the European H. arctica and the Siberian H. amoena than to H. parenthesis, and is, I believe, the Amer- ican representative of H. arctica. HIPPODAMIA. 55 3051 b. Hippodamia lengi, new species. Distribution: Southern California. In the Ulke collection in the Carnegie Museum at Pittsburg, Pennsyl- vania, there are 8 beetles which aside from color-pattern resemble H. apicalis. They are from California and bear a label ' Var. of H. paren- thesis Horn dix. " There is one other of the same kind in the U. S. National Museum from San Diego, California, and 'one from Southern Cal- ifornia in the Horn collection of the Philadelphia Academy of Natural Sciences. These 10 specimens are decidedly different in their color-pattern from all specimens of H. parenthesis or apicalis that I have seen and should, I believe, be recognized as a species, although it is to be considered as a derivative of H. apicalis. I have designated one of the Pittsburg series as the type. The pronotum is more extensively pigmented than in H. parenthesis or apicalis. In one specimen the mesocaudal light spot is wholly absent. In the other cases it is small and triangular, with the point caudal. The spots 4, 5, and 6 are confluent, sometimes leaving an inclosed light spot, sometimes with the pigment continuous. The entire suture is pigmented and confluent with the spots just mentioned. The pigment extends laterad in the basal region to spots 1 and 3. The pigment generally involves the base of the elytron, but in one case is free from it. Var. A. The spot 3 free. Var. B. The spots 1 and 3 free. While the amount of confluence at the base of elytron is thus seen to be variable, the distinctive feature of the species, the pigmented suture, is constant. The species is named in honor of Mr. C. W. Leng, student of this genus. 3052. Hippodamia faicigera Crotch. Distribution: Hudson Bay, Great Slave Lake. The specimens in the Horn and Leconte collections, in addition to their lack of discal spots and to the more elongate elytra, have a lateral margin which is straight or even slightly concave for a short distance. The vittse are much less sinuate in the apical region than those of H. sinuata (see fig. 7). A large series will throw light upon its relationships. For the present we must conclude that it has probably developed along an inde- pendent line rather than from H. sinuata. 56 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. TRIBE COCCINELLINI. Genus NEOHARMONIA Casey. 3055 a. Neoharmonia venusta Mulsant (fig. 44). Distribution : Indiana to Kansas, Michigan to Texas. Var. A Mulsant. The humeral spot united to the lateral spot of the intermediate row. Var. B. As in fig. 44, 6. Var. C. As in fig. 44, c. The pronotal spots confluent. FIG. \\.-n to c, variation of color-pattern of N. venustn. c, from Louisiana; d, e, N. notultitn. 3055 b. Neoharmonia notulata Mulsant (fig. 44). Distribution: Louisiana. Unfortunately I am unable to throw light on the interesting question of the nature of this melanic derivative of the previous species. It is here given specific rank with the greatest hesitancy. Typical pattern as in fig. 44, d. Var. C Mulsant. With the light mark reduced and divided into two smaller ones. The lateral one may not reach the margin. Var. D. As in fig. 44, e. 3055 c. Neoharmonia ampla Mulsant. Distribution: Texas and Mexico. The pronotal marks fused to make 2 longitudinal vittae (fig. 45). The elytral pattern (fig. 46) is much reduced, revealing in some speci- mens 4 separate SDots instead of 2 in the intermediate row and 4 separate a FIG. 45.— Variation in pronotal pattern of If. anipla. All from Brownsville, Texas. FIG. 46— Variation in elytral pattern in JV. (i/nji/n. All from Brownsville, Texas. spots instead of 2 in the apical row. This species stands intermediate in elytral pattern between Neoharmonia venusta and Olla abdominalis and its variation is parallel in general with the latter. COCCINELLA. 57 1RIFASCIATA TRICUSPIF CALIFORNIA TRANSVERSOGUTTATA MONTICOLA DIFFICILIS Genus COCCINELLA Linnaeus. The species of this genus recognized in the United States are shown in fig. 47. Reference to species outside of the United States is neces- sarily omitted on account of the large size of the genus. The species within the subgenera are closely allied to each other, while the subgenera are far less so. 3056 a. Coccinella perplexa Mulsant. Distribution: North Europe, Siberia, and North America. Our American species corre- spond closely to the Coccinella tri- fasciata of Eu- rope and Siberia. It differs in hav- ing the black area of the pronotum sharply reentrant, in most cases, at either side of the mesal portion (fig. 48). It is given specific rank with some hesi- tancy. A comparison of a large series of each would be desirable. Var. eugenii Mulsant. Cephalic band broken '(fig. 49). Var. == Coccinella trifasciata var. fennica Weise. Intermediate band broken. Form A. Intermediate band only missing (fig. 49, d). Form B. Caudal band only missing (fig. 49, j). Form C. Cephalic band broken and intermediate band only missing (fig. 49, d). Form D. Cephalic band broken and caudal band only missing (fig. 49, w). Var. Juliana Mulsant. Intermediate and caudal bands missing (fig. 49, /). A specimen from Cypress Hills, Long Island, was probably transported there upon plants. Form E. Intermediate and caudal bands missing, cephalic broken (fig. 49, g.) Form F. Intermediate band only present. Oregon (fig. 49, o). Var. subversa Leconte. Elytra without pigment. Pacific States (fig. 49, h). Form G. Black area of pronotum extends to the cephalic edge broadly as in Coc- cinella monticola. Rare. One from Berkeley, California (fig. 49 rf)., SUTUFWU3 PROLONGATA NOVEM-NOTATA JOHNSONI FIG. 47. — The species of Coccinelltt in the United States. 58 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. A series collected at random in the botanic garden of the University of California, Berkeley, gave— Var. subversa .............. . ............... ...................................... 1 Var. eugenii ........................................................................ 7 Var. eugenii transition to var. Juliana ................................ 8 Var. Juliana ....................................................................... 78 Var. Juliana, transition to Form C ...................................... 6 Form B .............................................................................. 5 Form A ............................................................................. 1 Var. Juliana transition to var. perplexa ............................. 2 Typical .............................................................................. 2 Large numbers show that the variety eucjenii is not as markedly set off from the variety Juliana as has been stated. This is indicated by the intergrades noted above. a d Flu. 48. — Variation in pronotal pattern in C. &=mode. The relationship to C. novemnotata suggested by Mulsant is not close, for the larvas offer a sharp contrast in color. The very closely allied European species C. trifasciata shows only one recorded variety (variety fennica) in which the intermediate band is broken. This variety was seen only once in one of the heredity experiments referred to later. In eastern North America C. perplexa is remarkably constant, but in the West it becomes more and more variable, always, however, in the direc- tion of reduction of pigment. Subspecies of Oregon characterized by the presence of variety subversa andF. Subspecies of California characterized by the presence of Juliana. Subspecies of Eastern States: The three bands seldom suffer reduction. Two mated females kindly sent me by Professor Terhune, from Lon- don, Ontario, gave me two fraternities. In one of these all 11 individuals were normal like the parent. In the other fraternity 1 of the 3 individ- uals corresponded to the variety fennica of C. trifasciata which I have not seen in C. perplexa. COCCINELLA. 59 3062. Coccinella tricuspis Kirby. Distribution: Canada and neighboring States. The elytral pattern of this species is so unique that its relationship is puzzling (fig. 47). The few specimens known show no sU nificant varia- tion. 3058. Coccinella novemnotata Herbst. Distribution: North America. Var. spots H-\-L, conjuncta Fitch. 4 per cent at Cold Spring Harbor (see fig. 52). Var. spots H-\-I. Var. spots I-\-A. Var. spots &+/. Seen on one side only of 1 specimen from Oyster Bay, Long Island. Var. spots L+H+7. Var. spots L -\-H -\-I-\- A, confluentn Fitch. Var. dirisicollis Fitch. Black pattern of pronotum touching the cephalic margin caudad to the eyes, but not mesad. Known only from Fitch's description. The prono- tum is in general relatively con- stant (fig. 50). Var. degener Casey. The pigment re- duced, the apical spot being re- solved into two (fig. 51). Prob- ably the specimens of C. mene- triesi Mulsant, attributed to Cali- fornia, are of this variety. But as Crotch believes that C. mene- triesi is a variety of C. undecim- punctata, the relation of that species to C. novemnotata requires further study. Form A. Some spots lacking. Var. jranciscana Mulsant. without spots. Elytra FIG. 49.— Variation of elytral pattern in C. Subspecies of the Plateau States: Variety degener prevails. Subspecies oj the Pacific Coast: Varieties degener and franciscana prevail. 60 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES- MODIFICATION AND HEREDITY. The application of 40° C. to the prepupae and pupae did not affect the pattern. The results of the experiments in heredity with this species are summarized in table 17. Since on Long Island the variety conjuncta was present in a ratio of 4 per cent, and each of the other varieties was found in less than 0.1 per cent, the mates of females from nature were probably normal. The progeny of such a female (278) are represented in fig. 52. They are remarkable, since 40 per cent of the prog- eny surpass the unusually pigmented mother in the amount of confluence of spots. It is hard to believe that this is;not determinate. The lack of intergrades between the 4 normal specimens and those like the mother is striking, especially in comparison with the experiment of 440 ?, and with the presence of such intergrades in nature. In general, the conclusion is TABLE 17.— Heredity in Coccinella novemnotata. FIG. 50.— Variation in proimtal pattern in CoccineUn not'eiimo- tata. 6= mode. Father. Mother. No. Nor. h+l htr.l h tr. /. h tr. in h+l, h t r. m ft f /-*-?)) No. Elytral pattern. No. Elytral pattern. 278? h+l+m 20 4 0 0 0 0 8 440 i + - h+l h tr. in 440 ? h+l+m 55 23 3 0 0 1H 4 405 t + — h+l h tr. m 40j ? nor. 46 39 .3 0 0 0 0 *449i »i + a '449 9 •in tr. i tr. /« 405? nor. 46 0 0 0 0 0 2 *449? ;/i -j- (f 1 111 i in tr. a 14 0 1 0 0 2 5 — 465 . h a p. 1 ft ap. in 21 0 0 0 0 0 0 Total 159 8 1 4 1 3 7 f- 'Progeny of 405. justified that there are here a number of positions of organic stability, namely, h + l, /i+l-j-ra, m-(-a, /i + l + m+a; hi- m was never seen without other confluence either in -these experiments or wild. In progeny from 465 9 (fig. 53) we see that an unusual shape of the humeral spot is inherit- able. The amount and kind of intergradation varies in fraternities from different parents, and from this we may infer that we are not dealing with sharp-cut unit-characters, but rather with positions of organic stability not finelv localized. COCCINELLA. 61 3058 a. Coccinella johnsoni Casey. Distribution: San Diego, California. This species (fig. 47) is obviously a derivative of C. novemnotata. Since C. novemnotata has the pronotal black free from the cephalic margin throughout its range and throughout all of its variations from the lightest variety, franciscana, to the darkest variety, confluenta (excepting the very rare and peculiar variety divisicollis} , it is probably entitled to specific rank. No typical specimen of the Western subspecies of C. novemnotata has been collected in Southern California to my knowledge. A large series is necessary to fix its status more definitely. a FIG. ol. — Variation of elytra! pattern of C. novemnvtntn in tlie Western State.*. /=mode. 96 °f. 40 % FIG. 52. — Pedigree from 278 •;• . 3059. Coccinella transversoguttata Faldermann. Distribution: Europe, Siberia, and North America. Typical pattern: With a lateral spot. (See fig. 54, g.) Var. quinquenotata Kirby. The lateral spot absent (fig. 54, d, e). Var. nugatoria Mulsant. The basal band broken. Common in the Western States and Mexico (fig. 54, c). Form A. Lacking humeral spot (fig. 54, a). Very rare. Seen only in Eastern Wash- ington. It is an extreme fluctuation of the reduction of pigment of the basal band characteristic of that'region. 62 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. The typical pattern is rare with us. Nearly all of our specimens in North America are of the variety quinquenotata. The variety nugatoria is common in the Western States, but not in the Eastern. Mulsant states that it is found more often in the south of Europe than in the north. Further knowledge of this distribution would be desirable. The extreme of pigment reduction shown is the form A from Washington. This is in good accord with the general belief that C. californica is a spotless deri- vative of C. transversoguttata and favors the idea of their close relation- ship. The variation in the pronotum is shown in fig. 55. e f 9 h FIG. 53.— Offspring of 405 ?. n> '•' I [<;. ~>l. — Variation ot'elytral pattern off. transversoffut/ala. tor, from Kamloops, British Columbia; /, fromStony Lake, Michigan; «/. I'roni \Vollaston. Massachusetts. 3059 a. Coccinella californica Mannerheim. Distribution: Pacific States. The small, yellowish basal area near the scutellar spot is present in greater or less degree, the variation being continuous. The suture is pig- mented ordinarily in California (see fig. 56), but in some cases approaches variety nevadica. Var. nevadica Casey. The suture not pigmented back of the scutellar mark. Reno, Nevada. Var. melanocollis n. var. Pronotum entirely black (fig. 57) ; 2 specimens from Berkeley. Two round white spots on the head found in this species and C. trans- versoguttata make it not unlikely that it is a derivative of that species. Unfortunately I was not able to make this cross. C. californica males were mated with both C. monticola and nouemnotata and found to be intersterile. COCCINELLA. 63 3060. Coccinella monticola Mulsant. Distribution: New England to the Pacific. In degree of confluence of the elytral spots Coccinella monticola (fig. 58) gives the following varieties which have pronota as shown in fig. 59: Var. sellica n. var. \ f median band. Califor- nia and New Hampshire. (Fig. 58, d. ) Var. postwa n. var. Median band -(- apical band. California. (Fig. 58, c.) Var. conftuenta n. var. \ - median band -f api- cal band. California. (Fig. 58, e). Fl(;- 55. -Variation ..r pron..i:li pattern in C. transversoguttata. Mode atob. The following patterns (fig. 60), although they appear at first glance very different from variety monticola, are nevertheless placed here, with some reservation, because of the intergradation from C. monticola in the direction of reduction of pigment in some localities. In any one locality the amount of pigment is tolerably constant. It seems to me probable that this amount in C. monticola is a result of environmental influence. The variation in their pronota is given in fig. 61. FI(T. oli. — Variation ol' elytra I patt'Tii of C. <'rni<'- mode. FIG. 57.— Variation in pronotal pattern in C. ntfifornicft. l> mode; ,•/, variety iiie/nii»cii//in . FIG. oS.— Variation in elytral pattern in C. nmn/ico/n. b -mode. FIG. 59.— Variation of pronotal pattern in C. mnntirvla. 6=mode. Var. alutacea Casey. Pigment of median band so reduced as to result in complete or partial isolation of a lateral spot. Colorado and westward. (Fig. 60, e, d.) Aberration impressa Casey. Similar, with "an impression along the side margin of the pronotum extending arcuately inward just before the middle, disappearing from the edge." This unique specimen is probably an aberration. The type specimen may never be duplicated, or at least may not become established. California. Var. sonorica Casey. The median band much reduced with the lateral spot absent. Colonia Garcia, Chihuahua, Mexico. (Fig. 60, b.) Var. bigiittata. Pigment much reduced. Lateral and apical spots absent. Buena Vista, Colorado. (Fig. 60, a.) 64 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. 3061 a. Coccinella difficilis Crotch. Distribution: Utah. Although the pattern of this species (fig. 47) is very close to that of C. monticola var. alutacea, it is probably entitled to specific rank. This is because of its small size and the fact that the two elements of the median row of elytral spots vary in the direction of greater pigmentation by increase in general dimensions rather than by confluence. Specimens are as yet too few to make its status and relationship clear. ft-fr. 60 FIG. 60. — Variation of elytral pattern in C. monticola var. nliitacea and FIG. 61.— Pronotal pattern of C. monticola var. nttttacen. n, n .single specimen. 3061 b. Coccinella suturalis Casey. Distribution: Colorado. This is very clearly a derivative from C. monticola var. alutacea by the addition of the pigmented suture— a repertoire variation. Its specific rank rests largely upon the reduced height of the elytra. FIG. li.!. — Variation in color-pattern in C1. protonr/nta. it, b, c, from Cheney, Wash.; rf, from Utah. 3061. Coccinella prolongata Crotch. Distribution: Kansas to Pacific Coast. The elytral pattern is essentially that of C. monticola var. alutacea, from which it is probably derived, but with the pigment still further reduced. The pronotal pattern is striking, and although there is some in- tergradation, it is apparently interrupted (fig. 62). The essential char- acteristic is not only the prolongation of the white area of the pronotum, but the spindle shape of the central portion of the back area cephalad as well as caudad. Its nearest allies have the black pattern extending along the cephalic edge for half its width. CYCLONEDA. 65 Genus CYCLONEDA Crotch. The color pattern and its lines of variation in this genus are markedly different from Coccinella, of which Leng makes it a subgenus. The form, agility, and the difference of the food of the larvae (Johnson, 1907) sup- port the generic distinction. The sexual difference alleged to exist in the pronotal pattern, though noticeable, is much less than indicated by Mulsant and Dimmock. 3064. Cycloneda sanguinea Linnaeus. Distribution: Cuba, Bahamas, Florida to California. Var. immaciilata Fabricius. With pale spots on pronotum inclosed. Florida to Louisiana- The sexual differences in the median emargination of the black pronotal area is either absent or slight. a FIG. 63.— Variation in pronotal pattern of Cyclonrdit male, n. It, from Cold Sprint? Harbor, New York; <: icta. , from Palmetz, Arizona; c, from Whitefish Point, Lake Superior; ii.<-itlvin duode<-iiiinin<-ui,it«: rf -mode; a, from Oregon: i>, from Hudson Bay. .Inixucn/rin quatuordedmffuttata: h=modc; << , and 0 in Epilachna b<>i-<-i/nchnii horeii/ix exposed to 40° C. FIG. 8t). — £,j>ilii<-lin(i foiri'H. n, comparative outlines of elytra from the side: solid line, K. Itnrrri; dotted line, E h/irenUx; l>, color- pattern of pronotum and elytra. 80 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. hb FIG. 9(1.— Varieties of //(/">>•dainia 8a !ia 13.8 2li.9 21.3 11.5 5.3 2.3 .9 3 .3 .4 .4 0.3 il 4 .3 .8 S.I 11.1 33.3 22.2 12.8 fl.9 3.2 1.4 .5 .7 .8 g 1.0 0.7 S QQ 5 (i .1 .7 .3 2.2 Ml 7 s 5.3 27 9 19.2 20.9 17.0 14.1 12.8 8.1 8.5 4.3 4.9 2.0 2.7 .9 1.2 1.3 1.7 1.7 3.2 2.0 4.3 2.2 4.8 2.6 5.3 1.7 6.4 1.1 1 ^ 7 .li 3.5 11. li 10.4 9.3 7.0 4.62.13 1.4 2.11 l.ii 7.0 9 :; 10.4 11.6 3.5 0.6 QQ S 1.1 (i.l 5.3 4 S 4.3 3.2 1 7 1.22.7 4.9 8 5 12.8 17.0 19.2 5.3 1.00.3 ^3 9 1.7 2.6 2.2 2.0 1.7 1.3 -92.0 4.3 8.1 14.1 27.9 7.812.2 .7 0.1 cc ft 10 .7 1.0 .8 .8 .7 .5 1.4 3.2 6.S 12.8 22.233.3 11.13 1 .8 .3 Successive 11 12 .4 .1 A .2 1 .3 .1 1 .1 1 .9 c 2 2.3 1.5 1 0 5.3 4.1 11.5 9.5 7.8 21.3 20 5 18.1 36.9 37.9 39.2 i:>, s 16.4 18.1 4.61.3 (i.22.1 7.82.9 .1 .6 1.0 0.1 .1 .2 0.1 .1 0.1- 0.1- -Nmnlior reduced to The evolution of varieties and subspecies is readily enough accomplished in these beetles. But the passage of the variety or subspecies into a spe- cies, that is, the acquisition of intersterility, seems much more difficult of accomplishment. The hypothesis of Romanes, that it has arisen by the endowment of certain individuals with mutual fertility but with interster- ility with the rest of the species, seems to me extremely unlikely, as such individuals would have very low fecundity from the slim chance of finding their interfertile mates. Of course there are a few exceptions in cases of particular causes of endogamy. A more probable hypothesis is that of correlation of some genital or developmental feature with the varietal features, so that they become sufficiently unlike to cause intersterility. There is a partial intersterility of Hippodamia convergent var. caseyi with 102 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. the species, and since var. caseyi has arisen within the range of H. convergens, this seems to be the most probable explanation, especially since the color-pattern difference is correlated with a difference in size, as shown in table 10. Where isolation can be resorted to, an explanation is simple. Determin- ate variation in the isolated groups has worked in different directions, so that the reproductive organs or processes may have often become so diverse as to make them intersterile if a breakdown of the isolation brings them to the test. In this I would suggest we find a reason why great organic differences in artificial varieties do not bring about intersterility, while much slighter differences in nature do. In the latter case determinate variation has had its opportunity; in the former it has not. In these beetles, however, isolation can only rarely give this assistance. SUMMARY OF CONCLUSIONS. (1) Variation. — Both continuous and discontinuous variations are found in the color-pattern of these beetles. Variations are also found disposed around certain centers of variation in greater numbers. Yet these centers lack the discreteness necessary to constitute them unit-characters. (2) Modification. — The color-pattern is capable of modification by the environment. Some modifications exist as hereditary characters also, while others do not. Non-hereditary modifications are more controlled by the structure than are the hereditary variations. The germ-plasm and the soma are in some cases capable of parallel modification, thus producing an apparent inheritance of somatogenic characters. (3) Distribution.— The species overlap to a great extent. The varieties occur with the typical species in a part of their range. Jordan's law is in general not followed; hence the evolution is probably for the most part not by the even flow of all the individuals in a region. (4) Heredity. — Segregate (alternative) heredity is general, but it varies by degrees from blending to a typical segregate heredity. Mendelian interpretations meet with difficulties in most cases. It is probable that we have preponderance (prepotency of characters) in some cases. (5) Phytogeny.- -There does not seem to be adequate ground for postu- lating a definite single pattern as the primitive one for the family. Eimer's laws of pattern development are not applicable. Several congeries of vari- eties or species of diverse patterns may be attributed to descent from a spotted pattern. (6) Evolution. —Natural selection, if at all active, is principally conserva- tive of the spotted pattern. In spite of this, determinate variation, largely actuated by the effect of the environment on the germ-plasm, and probably preponderance as well, have accomplished marked evolution of the pattern from this condition. Evolution proceeds by waves as well as by even flow and bv mutation in different characteristics at different times. BIBLIOGRAPHY. BURGESS, A. F. 1898. An abnormal Coccinellid. U. S. Dept. of Agriculture, Div. of Ento. Bull 17, n. s., pp. 59-61. 1903. Economic notes on the family Coccinellidae. U. S. Dept. of Agriculture, Div. of Ento. Bull 40, n. s., pp. 25-32. BUMPUS, H. C. 1899. The elimination of the unfit as illustrated by the introduced sparrow, /-V/.s-.srr domesticiis. Biological lectures delivered at the Marine Biological Labora- tory at Wood's Holl, in the summer sessions of 1897 and 1898, pp. 209-226. CROTCH, G. R. 1873. Revision of the Coccinellidaa of the United States. Trans. Am. Ent. Soc., vol. 4, pp. 363-382. 1874. A revision of the Coleopterous family Coccinellidae, London. COCKERELL, T. D. A. 1898. The zonal distribution of Coleoptera. Bull. 28, Agr. Exp. Sta., MesillaPark, N. Mex. 1908. Fossil Aphididse from Florissant, Colo. Nature, vol. 78, pp. 318-:519. CASEY, T. L. 1899. A revision of the American Coccinellidce. Jour. N. Y. Ent. Soc., vol. vn., pp. 70-163. CRAMPTON, H. E. 1904. Experimental and statistical studies upon Lepidoptera. I. Variation and elimination in Philosamia cijnthia. Biometrika, vol. 3, pp. 113-130. DAVENPORT, C. B. 1904. Wonder-horses and Mendelism. Science, xix., pp. 151-153. 1904. Statistical methods with special reference to biological variation. Second edition. New York. 1906. Inheritance in poultry. Carnegie Institution of Washington Publication No. 52. 1908. Inheritance in canaries. Carnegie Institution of Washington Publication No. 95. DE VRIES, H. 1901-1903. Die Mutationstheorie. Leipzig. DIMMOCK, G. W. 1906. Algunas Coccinellidae de Cuba. Primer InformeAnual de la Estacion Central Agronomica de Cuba, pp. 287-392. FITCH, A. 1861. Sixth report on the noxious, beneficial, and other insects of the State of New York. Albany. FLETCHER, F. 1908. Wild Jamaica cotton. Science, xxvin, 307-308. GULICK, J. T. 1905. Evolution, racial and habitudinal. Carnegie Institution of Washington Pub- lication No. 25. HARDY, G. H. 1908. Mendelian proportions in a mixed population. Science, n. s., vol. 28, pp. 49. JUDD, S. D. 1899. Efficiency of some protective adaptations in securing insects from birds. Am. Nat., vol. 33, 461-484. JOHNSON, R. H. 1907. Economic notes on aphids and coccinellids. Ento. News, vol. xvm. , pp. 171-174. KELLOGG, V. L., and R. G. BELL. 1904. Studies of variation in insects. Proc. Wash. Acacl. Sci., vol. VI., pp. 203-332. KELLOGG, VERNON L. 1908. Inheritance in silkworms, I. Leland Stanford Junior Univ. Publication. Uni- versity Series No. 1. LENG, C. W. 1903a. Notes on Coccinellidae. I. Hippodamia. Journ. N. Y. Ent. Soc., vol. xi, pp. 35-47. 19036. Notes on Coccinellidae. II. Coccinellini. Journ. N. Y. Ent. Soc., vol. XI, pp. 193-213. 1908. Notes on Coccinellidae. II. Chilocorini. Journ. N. Y. Ent. Soc., vol. xvi., pp. 33-44. 103 104 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES. LEIGHTON, ROBERT. 1907. The new book of the dog. London. LUTZ, FRANK E. 1908a. The variation and correlation of the taxonomic characters of Gryllus. Carnegie Institution of Washington Publication No. 101. 1908?;. Notes on the inheritance of variations in the color-pattern of Cr/occr/'.s- asparagi. Psyche, vol. xv., pp. 50-52. 1908c. Combination of alternative and blending inheritance. Science, 28, pp. 317-318. MEISSNER, 0. 1907. Die relative Haufigkeit der varietaten von Adalia bipiinctata L. in Potsdam (1906), nebst brologischen Bemerkungen uber diese und einige andere Coccinelliden. Z. f. wiss. Insektenbiologie, Band III., pp. 12-20, 39-45. 1908. Die relative Haufigkeit der varietaten von Adalia bipiinctata L. in Potsdam (1907), nebst biologischen Bemerkungen iiber diese und einige andere Coccinelliden. Z. f. wiss. Insektenbiologie, Band III., pp. 309-313, 334- 344. MULSANT, E. 1846. Histoire Naturelle des Coleopteres de France. Sulcicolles et Securipalpes. Paris. 1850. Species des Coleopteres trimeres'securipalpes. Paris. 1866. Monographic des Coccinellides. I11 Partre. Coccinelliens. Paris. MATTHEWS, W. D. 1908. Osteology and Phylogeny of the American.Cervidae. Bull. Am. Mus. of Nat. Hist., vol. xxiv, pp. 535-562. MACDOUGAL, D. T., A. M. VAIL, and G. H. SHLJLL. 1907. Mutations, variations, and relationships of the Oenotheras. Carnegie Institu- tion of Washington Publication No. 81. POWERS, J. H. 1908. Morphological variation and its causes in Ambhjstoriia tigrinum. University studies. Pub. by Univ. of Neb., vol. vii., pp. 197. REICHERT, ALEX. 1896. Uber Coccinelliden und ihre Variation. Illus. Wochensch. Entomologie, Band I., pp. 26-30. 1904. Die Varietaten von Adalia bipunciata. Entom. Jahrb. fur 1904, pp. 179-181. RATHVEN, A. G. 1908. Variations and genetic relationships of the garter snakes. U. S. Nat. Mus. Bull. 61. SIMPSON, FRANCES. 1903. The book of the cat. London. SIIULL, G. H. 1907. Elementary species and Hybrids of Bursa. Science, n. s., 25, pp. 590-591. SCHROEDER, CHR. 1901-1902. Die Variabilitat der Adalia bipunctata L., gleichzeitig ein Beitrag zur Descendenz-theorie. Allg. Zeitschr. fur Entomologie, Band 6, pp. 355-360; Band 7, pp. 5-12, 37-43, 65-72. TAYLOR, JOHN W. 1894. Monograph of the land and fresh-water mollusca of the British Isles, pp. 386-390. TOWER, W. L. 1903. Development of the colors and color-patterns of Coleoptera, with observations upon the development of color in other orders of insects. Decennial Pub- lication, Univ. of Chicago, vol x. 1906. An 'investigation of evolution in chrysomelid beetles of the Genus Leptino- tarxa. Carnegie Institution of Washington Publication No. 48. WEISE, J. 1879. Bestimmings-Tabellen der europaischen Coleoptera. II. Coccinellida?. Zeit. fur Ento., Breslau, Hef. 7, pp. 88-156. NOTE: This paper was completed October 3, 1908, before the appearance of Casey (1908), the result 'of which could not therefore be considered, except for certain advance information kindly supplied by the author. CASEY, T. L. 1008. Notes on the Coccinellidas. Can. Ento., XL., 393-421. November, 1908. M.?!s WHOI LIBRARY UH Iflix N