UN. oF

ets Perse yao et Tartin a

L pes bask Sey

ae Sa ere { mame, CS ae.

‘http /www.archive.org details/determi nateevolu00johnuoft

~~

yaar ie

oF : 7 GP? eh: 4* PETERMINATE EVOLUTION IN THE COLOk- 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 issued

JUN 29 1910

THE CORNMAN PRINTING COMPANY, CARLISLE, PENNSYLVANIA,

CONTENTS.

Page

PART I—INTRODUCTORY CONSIDERATIONS. ...........02-2c00edeesesesscessenccccecssccecseeseecesscrers 3 Principles of classification...............-.-s--ssssccccrecsssessscceesecssenccesencceseceesseeeesen seers 4 Structure of pigmented areas.................--..-escee--sseenerenseeeseenceseseeenenseeneees coenes ~ 8 Development of pigmented areas..............-.2-ccscsecsesenereteeesenenennenneceeeseenenetenseeees 8 ICN RUN NOU INMEL ONIN. oon ccnitne cede acapiinnessoansonstisisibadduitp~osceqneessnatsenseasces 9 Nee en nee oe sea xnterscen AAAdapOep Renard csacdavsangeressacas 10 Variation and heredity of body-length......................2-:ssesecececcseeenenseeeeeeneeeeeeeseees 13 I pc ah eek wtb ccnp anis ba saiain sah a= ah4Od RAMAN ApRRN ss eoungnhedbouis 14 Part II.—THE DATA ARRANGED BY SPECIEG. .............--2sscescecesce-ceeceesntncseecccsseneeneenee 15 dal oe sc Peca hana ie tad gaa y esndb kananeomicenaaséanannaachbpsasbbdnieo=nesonesensnenes 15 i a can 5a conan eanbeanenedsaceocanqucanessdneebea pam | I irdceiceei tag Ririas toncnndienescnxdexepnainrsrese=n tmpeaares sdbbgpassasaves 15

atl aie Rika oa See RR cdc ded hops danccidacobsasaneahase dusnceniinGetanDigha osetisen 15 I ee ic asacaanthenadedasadpenns sai Reaeeeees™ 15 IIe a os app usaknnmnpansansatekoehialeteubctassedenseas 16 ng ESE SRE RENE SESE SEIRS SOB ee a enh ae 16

TEN ae eas obec ccnp bid hae ns bnnin con gulse ses gewbiedasenoucsasae cs 16 I enacts oh olla icuiniapchishajaneia Site ns'aenoninldexhoanen wa eee 16 Tn osu uhaudh wud asuavedenvanentAiGhbbidsusdaunsesn 16 NU So apa tea ans sansa ccancncvauesudey inesoagends-ceaSanksaataniienneeraaek 16 Lass cinan cnerannndinnannnnnnmonciindnatnnndenankeaheeeinieiees eRe Es AIS SPARS et De be ee Eee eae ee ee Bees 17 |

Ne hah ie nUM GEN Salesedigachiduicienta wan penansninhian as <n jasptandrhaanennnnnhihWidheeataktaiee= ee | meet Mabe ep Sis a! ae Meee ep ease Gren ae er nS Pes tae ere ede ai ee

PS eee ae eee cee pelts eck sactteiepitetsanccaces Sekcceeves 17 ET iia ea met oS «Se Vite 5 ne A aE ot eR Pe 2 19 convergens.......2.. EM GG 85 nah ailonn siting mens hci pantieabbeninsonaccustnesa 21

bowditchi ............ ileal shbeebbcektnpidhcnasdsscpnsopancndpeaapbetaabessnetniates = Caen 6 ROE SES CE eae on" os ee nee Aer 46

IIE thlodcketdvpinbhewahisinkstesess Sha peaincednusne sedeiilekuile oascansnave - 48

SR eA, Se al ge SE . 49

ita ar See SAS 2 7 a Sipe ee en, 50

I io enwasanenecoeneece 50 PI cv iveiinewsbavsnevcions - 50 a sw scncdbceunsane 52

po SSR RSS oS ee <r er Te

ea entare te TRESS 5 SERIA. ee ae ee ee Pe Re . 5

WN oa ne easis pe nene pena Aes bh Sts perasieseninvisstonnrareesy sedecess baad ss 54

ee EE RPEIRER LSE a a 55

ON certain tt iiracsana'ncxnncyae neat cca eaiacconcic cs bobecdoccecersseeedsbaadathised fuse 56 De cn ornssi<annence beech ractieverae os enadsécien nes sccecdacceal ¥seoteacbieces 56 SRS Se ES ey ere ee 56

SR soca aca tine reine aga ts ohn onacicnnnioocuscbueadestcs cade <0 56

NI. 5 eiinns gant insinn Vapabaeet vas ubsaadtbbass seseveshesttilicnddipaadtinn Sim vcchnwases . 56

iv CONTENTS.

ParT II.—THE DATA ARRANGED BY SPECIES—Continued.

Coccinellini—Continued. Page Coccdinelia.s..6icjiicds sstgida estan ss S cd gobidtevissivc nae Meee ECAR UME MEE os ca cee 57 PEPPER .... «0... sds insqsped daapacapngns<ienbs see eeeN eMEN Neate Rene eee eam eee 57

tHICUBPIS ........5000...0insababnnganse ddbtn bay Sanpn es mateagtele kaa haes eer aeeenannn Ee 59

noOvemnotata ... ...<ccercscadsaseqnessuvcnsscungoeeduneoeeeetauadense: dusk aeneetaeeeeeaes 59

JOhNGONL, .....000: cascode alvasesa tien sMedeoesnabcapaeen alpaaeent ees aaetaneaeen 61 transversoguttata ..........-ccntuedsessbenbaligietuds otcteverwtaeay petal acid actcek 61

CALIPORDICA «0.055: 50cn00sxcacdenseracencaiecusntteaie idemaditgan Tee eek 62

MONTICOIA ...... +000 s0veeveingsns vend gnedsseeeMMvule ss Poaglebis tyeas palewi ae ean 63

Gifficilis. ......3.<scseccssnay Wiese vue vasnledeeenee edi oadetile wana eine Seema rears ay, |!

eeaterrall LG. « «is 500 caseedapns none dys <anks cee dscphaRieneecae Suan hy cask» anes dl ane 64

PPOLON GAL .0ssseserseeeseviedaseasivass Urbudud dunpeannsecsapesoadel lads ela tees 64

Cy cloned. ........cssvsassusaceegsevenucndors sevitendedadeeia ce tdgsten bray seesane bmn Uae aiane SES cca 65 BAD QUINOA. sieve cases evvnsecsesseedivuvovedncseuduentiupecabedenundeaae iene naan nies 65

AUN 645.s8ssssieasssc0s vices cine Ovees ceeenes cocoa nee ebeeaate ded acetate anes (ha dints whee 65

QLED. vevcecdsvecvacvawsovesdesewsvelbwdeubewsevdesscewevuoes/ssaasive’s ote lenis intent Cane aIE REE: 65

Old. 12.20. eccscvesvawevs reversadesucsseweuseueucedavwevewewtoanCecsereneeel pameuale ees nan mee 66 Mbeominalig ociessicessiacesscsdcccsets vevarenbwavs coventeseak acute vale aneieIne EMER pammnaE osea aa 66 Pha GI AEA v.0ncewidcessepannsoasscessbuecqrceveeaisenss0eo¥ponteieihon Ss msLeae hy deal ie DnEsesneeee 66 Adalié cccscacctacess cisnuneesdeoh iin o}'b ew. undabgdi 0a Sonus eae ee knee alee mERs setae econ oars 67 Dipamctata. | .ssississeccéssesscovvescaksansechseus ices eieebermdenssaenc acelin eeeNnN te. cea cueled 67 EVIGIGR bis. 0 sec ccp eaves vin cense nsdueccnnnebelaadeanyson NeoPM PMGEEE ccsvncs sats i useubsuacien eae 68 QNNECCADG 605. scsechenscddacssnnknsWscaehes but hee phaseeeeeres Unevu ein shcsak bee sUey eta ORCEn een eanees 70

Geis i. ixixisces duds vaeadae sawed on yeceneecuunceeeeteaekaie celana ny aneeecd ane RY Sica cipucdeundaten onneeeen 72 DICER paccieewese. Laas enwabievs oe srap diy stipe DR outa a lanetaindicas & Rinlbaatan! puebahcdeo hiner RED ta eee tiles 72 WUASONI CB e555 5 Se Snsder ce unsecule sud vay ioenanca duped tageed 4 Vawse dest aaa eek kaa ae iia ee co The 72 Agrabia . ssicviisccasausousshnvebsvaveeyeesendeneeuona pene evnws Gab aibea bibadGsiiessaieiannnnnMMEINELE ova w cccs 73 CYANOPCOl A sien convccevwaverenns dvasiaapuoswacheed bets sane Rae Dees even atop siiy =p waeaamiaanee 73 Anisocal Via x s+000scwvcewssevs vecchseucOnconedn eh suse t eyes eine seiaieo Odie ame tie ae ee 73 duodecimmaculata. ........ yids ocahissip Evi eae ee Cad eC noses rok i eaane 73

qua tuordecim@uttatar «4c. s.isis52.0acauedenesesvevascancensyatePieueew elds bdescssecons 73

Amatis 5.0.5... icccesueeueeubenanlecvesutau sateen eva yas eure esteem eaten i I PE SEOES STS sh os wears neeic’s 74 quindecimpunctatar oisecssascncecsiesescdsosessnipus/) RMUEMRUENND vas orsiisinivysseceesessase 74

Mali.......tvgbssenscesses conchae bub epieabasrdccassdendeteaad Uo Ma aM EMEP GEES IRA Sia cZai'vws veneer 74

LACH VON. cic Feveaceec cerca dress Laem ue meme eR he aan See Whe WOR MENTE oe Nes tesiecianw'' 74 leconted=sisisccaisixiskoaed cde yents tad ab dabig tated sap Tee ne ete ODS «Cea O GA eaws sak Sinwey 74 Epilachnini. «....:s.e.sisoutasesegetunesd<oieie webbie re sQieeciin dears VOPTEN MME MENER Gens 02 5heScaEsEUs sdsive es exes 75 Epilachina «..<<<c.evevnstdipuetees Gpinesssitegis pete tangs arta tateev AAPM ely ices sbaneebsediesesesecss 15

Dor eeh isis ais caVorakees cas icas asad udacdseeenyeees ely Dee eeOEM cease aves cabtarscncescsves 75

COW OER ois week gs. spetuoa seve an ee eR er aE tedes cauares pais acesenses 78

CORTU PUG se cciesins cs cangeescs cu5i 04455437 scans semUNeatn ETERS BERS oij ncaa sates tcsecceee 79

INO@RICAIB Sc sasccacuecisnd svactidaceavakernyii RAReAtEUMEOEE EE Np yaGicesy daececsecacece 719

PART II].—GENERAL DISCUSBION, «2.25 0ccdsevccosabuoneeepedees wie NPUOEOEIMI sa secnce scccccnessecceccsce 81 Variation ..... 1.06: sssaseccdeabenteresssied G1g814 eet TUN aida ORC aases00s cveassccsesers 81 Modification. ....<cixss..<.vewseweecovadusedsulladecy tadeae tees eeaen ERIE Teas W055 cco vecediuacaennenes 84 Distribution. ..2.665.10cccsassecgunes pee een tobe ete Ala tenibakaseeth ssi setok ethan sce 86 PROreOdIty........00csccncessscssasdvadabace sed labeeadaracdtGigeieenmnany sa 121i 18ti ates Von Cee bie Sevicnenas 91 PU YLOC ODS oie vase cvevite cucu devveaelpevous avedersn ARUN dois Ts TOMER 3 cavSgcinateses: 94 BW ONDGI INA cies ss osc csecvndnnd badeusscos vesecevevcddsil neuen atatiord ie Hem IER che sn 09 feeb tine vase 95 SSUNMNOTY OL: CONCHISIOUS cco atiie doce ccasssoscerssnclesckdsddadevece act CMMEA Tia ts bne0 8s aNsadasescdeen 102

MI G50 a rape RMA Eas s «oss csaucdeucnaveupeeybocuncntt coe aie Patent MeRGMaNTRGIS Oaks aida yas saxesins 103

DETERMINATE EVOLUTION IN THE COLOR- PATTERN. OF THE LADY- BEETLES

BY

ROSWELL H. JOHNSON

PARE Lb: 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, Coccinellz, 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 mucl 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 Ainerican species within these groups is given for the sake of completeness, although for some species but little variation data was obtainable. Tais 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 ntroductory section. Furthermore, this arrangement will be more advartageous for those whose interest in this contribution is primarily taxononnic.

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 varieti 2s and species in these remarkable genera are so doubtful. Iam obliged, therefore, to present a revision of my own in order to have a suitable ncmenclature. 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 man; characteristics.

Not having studied the «tructural features of the foreign species, I have not ventured to revise the genera and have simply adopted the genera

as used by Casey. 3

4 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 larve 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 convergens 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, ap to indicate a close approach to confluence, s/+ for a very slight confluence, and tr for transition between separateness and confluence.

Formule 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 itis 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, ete. 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. Iam well aware that many of these terms areused 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. Atany 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’’ may be 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 much more subject to modification.

STRUCTURE OF PIGMENTED AREAS. 7

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 could 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 bya 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 convergens. 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. a b. i

Fria. 1,—Position of linea interna, linea media, and linea externa, of: a, d, Adalia; b, Coccinella; ¢, Hippodamia; d, Anatis; e, Epilachna.

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 Adalia 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. novemnotata 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 thesoft 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 out- lined by pigment in cases of modification by cold and in the two wild

8 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

specimens shown in figs. 25, 16, aa, and 16, bb. 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 vittze when 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 vitte.

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 4and5 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 istrue in all the species which I saw emerge, including Hpilachna 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 aseparate 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 E’pilachna 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, 7. 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 Coccinellidz we have showy insects, ill-scented or flavored, that are eaten by but very few birds—the fiycatchers and swallows; and hence here is a whole family which conforms well with the theory of warning coloration. * * * The showy, ill- flavored Coccinellide [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 versus 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 larvz, 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.

Epilachna 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 sueceeded 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. Megilla maculata hibernates, at least frequently, in masses, one of which was found and kindly sent me by Dr. Robert W. Hall, of Lehigh University. Iam 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- vergens and spuria resort 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 paper. 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 Cycloneda munda at Cold Spring Harbor. This coccid is not eaten by any of the other species. Megilla fuscilabris 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 Rhamnus cathartica apparently is attacked only by ‘Adalia bipunctata at Cold Spring Harbor. This may be partly attributable to its early and short season. Upon the willow, Adalia bipunctata, with only rarely a specimen of Harmonia picta, is found on Long Island. The preference of Coccinella monticola for the larch has been observed by Dr. G. W. Dimmock. At Cheney, Washington, the aphid of the box-elder is especially attacked by Adalia. Other preferences of coccinellids outside of the groups here treated, such as those of Chilocorus similis, Pentilia, and Vedalia, are well known in the literature of economic entomology. The species of Hippodamia and Coccinella, however, show fewer preferences and are therefore found in more general competition. _I have elsewhere published further data in regard to the food of lady- beetles. (Johnson, 1906.)

Epilachna, as is well known, differs in being solely a leaf-eater. In this genus the species differ more in their food, for FL. borealis eats the leaves of the cucurbits, while E. corrupta eats those of the bean.

Sexual attraction is especially powerful in the lady-beetles, since so large a proportion of them are seen in copulation. This is partly due to the prolonged time of copulation. But when the pairs are associated, the male nearly always attempts mating at once. Mismatings of species with spe- cies and even family with family have been recorded more often, I think, in the coccinellids than in any other family. This is significant in connection with sexual selection. for, where promiscuity prevails, as it does here, the chance of either preferential or assortative mating is decreased. A female may lay fertile eggs as long as 3 months after mating and possibly longer. The beetles are emerging at intervals throughout the summer, the broods overlapping. Promiscuity is, therefore, an advantage in this species, for if the male did not mate or attempt to mate frequently, many of the females would lay batches of sterile eggs before being fertilized, as this is their habit when reared and not provided with mates.

The eggs are laid in successive batches of 1 to 40, most frequently about 20, at intervals of a few days for an extended period. The competition between the larve of one batch is frequently very severe, because the existence of the colony of aphids which excited the female to lay her eggs is vicissitudinous. Aphids suffer from attacks of lace-wing flies, syrphids, and diseases. The latter two causes of death are especially serious and frequently destroy a whole colony of aphids upon which a fraternity of

12 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

lady-beetle larvze are feeding. These larve then disperse in search of another colony of aphids. In this they are rarely successful, for colonies of that species of aphid in the immediate neighborhood are subject to simi- lar attacks and with the larva’s limited capacity for wandering the chance of finding another species of aphidsis indeed slight. As a consequence, more larvee would be successfully reared if the females laid their eggs singly in good aphid territory, for large batches of suitable aphids are frequently seen without lady-beetle eggs or larvee. Any change in this direction is not possible, however, because there are no favorable variations by which the lady-beetles could locate a sufficient number of colonies to distribute their eges properly.

The less of two evils, then, is to risk overcrowding and the consequent failure of some of the batches. We ordinarily think of overcrowding as having no other result than the desirable one of the survival of the fittest, but this other result of group-suicide may and frequently does follow. I have often seen a batch of eggs laid on a plant where I could see that the colony of aphids could not last through the larval life of the coccinellids because of syrphids or attacks on the aphids of disease. The larve would then disappear among surrounding plants destitute of aphids. In Epi- lachna, on the other hand, the laying of eggs in batches is advantageous, for shortage of food-supply is unlikely. In confinement the larvze, except those of E’pilachna, are cannibalistic when the aphids are exhausted. This is probably not an important feature in nature, however, because the larvze are so rapidly dispersed when the food is exhausted.

Although assortative mating is apparently not operative in these beetles, there is a very remarkable assortative association. At Oakland, California, the typical Hippodamia convergens and its spotless variety are both common, yet a hibernating mass taken by Mr. Nunenmacher and kindly given to me consists almost wholly of the spotted variety, as shown in table 6. If other hibernating masses are found to be similarly segregated we have a noteworthy condition. It is doubtful if this produces much or any passive assortative mating, however, as dispersive flights for food doubtless pre- cede mating in the spring.

VARIATION AND HEREDITY OF BODY-LENGTH. 13

VARIATION AND HEREDITY OF BODY-LENGTH.

The variation in size in some of the species is really extraordinary. It is relatively slight in Epilachna, Adalia, Harmonia, Anatis, Cycloneda, Paranaemia, Megilla, and Anisosticta. Butin Hippodamia and Coccinella H. convergens and C. novemnotata stand out as greatly more variable than their co-species (fig. 2). It is easy to see why Epilachna should not be so highly variable, since the food-supply rarely becomes short for the indi- vidual larva. The aphid supply is, on the other hand, very uncertain, and it is a great advantage for a species to undergo metamorphosis when the food falls short, even though undersized. But why are not all the aphidivorous species highly variable, then? Again we may say that Adalia is more constant because of its short larval period, which makes it less likely to suffer from limitation of the food-supply; but why, then, is the large Anatis

i i Ul i i

A

~ a -_" a a“ ~—— = ae, =e ~~ - a > ot

~ a -—

An /\\ / \ ‘ i \ \ ‘ x \ 68 69 70 71 72 7% 74 75 76 77 7S 79 80 81 82 83 SM & 8&6 87 71114419619 8i1200000000 016 368 60 16119149 8164129 3 5 2 0=1%

154 males ; 175 females.

Fig. 2.—Variation of length of elytra of Hippodamia convergens, Fairfield, Washington, 1905.

relatively constant? To determine if shortage of food-supply really could reduce the size, I experimented upon Epilachna, which one would expect would be most resistant to such treatment. I was able to get healthy beetles of only 5.7 mm. in length of elytra, while 8.2 mm. is not rare in nature. I believe, then, that the greater part of the variation in size is modificational. In addition to the factors mentioned above, it seems possible that H. convergens and C. novemnotata resort to a greater variety of aphids, the season of some of which is of short duration.

14 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

ACKNOWLEDGMENTS.

Before proceeding with the detailed discussion of the species, I wish to express my gratitude to the owners or custodians of the following collec- tions for the privilege of examining their beetles in these genera and in © some cases for the loan of material: United States National Museum; Philadelphia Academy of Natural Sciences; American Museum of Natural History; Brooklyn Institute of Arts and Sciences; Museum of Compara- tive Zoology, Cambridge, Massachusetts; Boston Society of Natural His- tory; Carnegie Museum of Pittsburg; University of Ohio; University of Missouri; University of Nevada (Prof, Peter Frandsen); F. Blanchard, Tyngsboro, Massachusetts; F. W. Bowditch, Brookline, Massachusetts; T. L. Casey, Washington, District of Columbia; W. T. Davis, New York; Charles Fuchs, Berkeley, California; L. H. Joutel, New York; C. W. Leng, New York; F. E. Lutz, Cold Spring Harbor, New York; W. S. Marshall, Madison, Wisconsin; F. W. Nunenmacher, Goldfield, Nevada; Carl Schaef- fer, Brooklyn, New York; E. C. Van Dyke, San Francisco, California. I am further indebted to many friends who have sent me beetles from various localities. To the Director of the Station, Dr..Charles B. Daven- port, my gratitude is due for my interest in evolutionary problems and for aid and encouragement in undertaking this one. I would also thank Dr. Schwarz, of the United States National Museum, for help with the Ger- — man citations, and Prof. W. L. Tower for suggestions upon the structure of the elytra.

Except where otherwise stated, it is the intention of the author to deposit the types in the United States National Museum.

a nd =

be Se ee ae

THE DATA ARRANGED BY SPECIES. TRIBE HIPPODAMINI.

Genus ANISOSTICTA Duponchel. 4033. Anisosticta strigata Thunberg.

Distribution: Northern States and Canada.

Var. dohriana Mulsant. The subapical spot dnly free from the pattern of confluent spots. Typical specimens rare in North America, but have been taken on Vancouver Island. Var. bitriangularis Say. Others than the apical spot free from the pattern of confluent spots. Most common

in America. Var. A.

Spots all separate. Manitoba; Sand Hills, Nebraska. These specimens probably represent a subspecies, for a specimen showing nearly as much separation came from Beaver Dam, Wisconsin. Specimens of the typical species of

the variety dohriana 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. irregularis with 1, 2, 3+ 4, 4+5, 6+ the suture, 7+ 8, 9 (his numbering), from Oregon. This throws some doubt on the specific distinction between A. novemdecimpunctata and strigata. The difference in the elytral markings are certainly bridged over by our variations.

Genus NAEMIA Mulsant. 3034. Naemia seriata Mulsant.

Distribution: Maritime regions of the eastern United States and south- ern California.

Form a. The spots separate, but the apical spot meets the suture.

Form 6. 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.

Subspecies of Florida:

Pigment reduced, so that the spots are not confluent. In this it is analogous to the

subspecies floridana of Megiila gpa er 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. Macronaemia 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 vittigera 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. similis.

a b c d e cx oh

Frag. 3.—Variation of the elytral pattern of Megilla maculata. c=mode; a to g from Shawnee, Pennsylvania; fA 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. a b c

Form 6. 1+ scutellar mark; 2 individuals (fig. 3, f) Fig. 4—Variation of the pronotal pattern

Form c. An extra spot between 1, 2, and the of Megilla maculata. b=mode; «a to e¢, scutellar mark; 1 individual (fig. 3, d); froma Ehawaee, Pennsytvania,

1 also seen at Cold Spring Harbor, New York.

Form d. 3+ 5; 1 individual (fig. 3, g); 1 also from Texas.

Form e. Spot 2 resolved into a larger internal and a smaller lateral spot (fig. 3, 6). 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 f. 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.

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 antenne 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.

Genus ERIOPSIS Mulsant. 3040. Eriopsis connexa Germar.

Distribution: Texas, California, and Vancou- ver Island. ee LN

A South American species rarely found in the *““ceratomeyia wien United States (fig. 6). It is possibly derived ag pe on gs hoon from a spotted Hippodamia by reversal of pat- ae 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.

SP SW CW UW COW? DB

wa WV TW WD

GLACIAUIS

BOWDITCHI

CONVERGENS

DISPAR

COCKERELLI

OREGONENSIS

SINUATA

SPURIA

TREDECIM-PUNCTATA

AMERICANA

PARENTHESIS

APICALIS

LENG!

FALCIGERA

Fie, 7.—The species of Hippodamia in the United States,

HIPPODAMIA.

TABLE 1.—The species of Hippodamia.

19

; Group. Species. Locality. _ (giacialis. ......< ere Northeastern States and Canada. Convergens. .......+..2-+0+ pwwehiteah: os. sac 4b ue se> Northern Rocky Mountains. COREVORMOIG 5 occ ccgecscecss North America. | feockerelli............-..2+++ Rocky Mountains. Co Se Se Sees ae Colorado, Spuria ......-...--+-.4--+++- oregonensis............-..+- Cascade Mountains. | | spuria......-....-2- +--+ eee Western States. | Usinuata ... 2-26. sees eee ee ee China and Pacific States. ide 3 | (variegata ..........0.0ee000- Septemmaculata variegata { Dae... BEPNTORER. [555 52S lec cccces } Tredeci tredecimpunctata ........-. Europe, Siberia, United States, Canada. sige obama RAIOTIOUINE Shc von dah os aeeees Hudson Bay, Lake Superior. parenthesis ............-..- .| North America. WECM oc icc e ea dpepnssaneses Western States. PMSUMINORS 2 oi coke soos 5s Nox See sos ustiee ee os- << = Southern California. SAINTE a <'n, «0 cig eacelh oes os le | Eastern Siberia. OR «oc ce ccnde eg bes sneune Europe. ee ee eee | Hudson Bay, Great Stone Lake. Incerte sedis. .............. {kreichbaini Ree | Abyssinia. © ’ S| SESE Ae See ee

TABLE 2.—Comparison of mean and extreme dimensions of some species of Hippodamia.

) (Measurements, except those in percentage, expressed in units of 0.0625 mm.]

Breadth | of | preaath | poreadth of [ 1 | Length Breadth | rightelytra | t ronotum In | Species, sex, and locality. No.| of Tient ofright | in percentof | of peene: oe cent of | elytra. elytra. | gee roe of um. ee | | | Hi, ronvergens, Cold Spring | | rr: ; ) Male |. ates Se a a7 66 76.9 8 25 30.6 36 | B15 40.1 47.5 | 36 41.1 44 48.5 57.4 59.5 H. glacialis, Goid Spring Har dot hacer mle | ead» air r: | : epee Sane 25 | 78 96.3 94 28 35.441 34.5 41.2 46.5 41 46.6 51 | 50.5 54.2 59.5 = ke gp ee wp ips 29 | 30 94.2 104 | 33 38.3 47 | 34.5 40.7 49.5 | 41 49.2 4 | 49.5 52.4 56 aeons "eee-| GL | 55 63.2 69 | 20.24.3238) 32.5 38.5 45.5 | 29 38.937 | 48.5 53.8 57.5 ” me sia take 75 | 54 68.2 73 | 212.431 82.5 38.7 44.5 29 35.5 BY | 47.5 52.1 66.5 | Pa Se ee 53 5659.9 64 19 24.229 | 32.5 40.5 46.5 | 30 33.7 37 | 51.5 56.3 60.5 | Wemnblbres: 552. DR % 53 62.8 68 20 25.0 30 | 32.5 40.0 47.5 | 30 33.8 38 | 50.5 54.1 61.5 | / | | j |

3044. Hippodamia glacialis Fabricius. Distribution: United States and Canada.

é The color-pattern (fig. 8) of this species is approached and even realized in some cases by specimens of H. convergens inthe 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 larvee, is such that it has probably been derived from that older type at no very distantdate. The pronotal pattern is like that of H. convergens, but shows relatively less variation (fig. 9).

Var. A. The humeral spot absent. In9outof 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. Spots4 + 5+ 6. Several taken from di- verse localities. Nowhere established (fig. 8, 7). :

Var. C. Spots 6+ 4 + 5. & Vea J As above (fig. 8, h). c

Var. D. Spots 4, 5, end6 | |

merged in one rounded area (fig. 8, 7). Only 1

e -

specimen seen. _ h Var..E.. With spot No. 2;

3 cases in 55, or 6 per cent, eZ

at Cold Spring Harbor ¢ | b

(fig. 8, a). | Var. F. With the extensa Se

mark. Only 1 specimen at Cold Spring Harbor

Q

(fig. 8, 9). Var. G. Spots 4, 5, 6 (fig. 8, d).

x ox eee Fie. 8.—Variation of elytral pattern of Hippodamia glacialis. Var. H. Without discal han A.

spots on pronotum. Only 1 specimen from Cold Spring Harbor. Also seen in the H. convergens var. pseudogla- cialis of the Western States.

Var. I. Median white spot upon head extending to eyes. Especially characteristic of western specimens (fig. 10).

BDRDRPD: 373%

FiG. 9.—Variation of pronotal pattern of Hippodamia glacialis. b= mode. Fia. 10.—Variation of color-pattern of head of Hippodamia glacialis. d=mode; a from Onaga, Kansas; e toe 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.

ea ia

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. ee 2 | 4and5

| No. E = Very | | | Sepa- |1, No. Pattern. } Present. | pmail. | Absent. | Present. | Absent. | vate: United, | 28 g 2482 | labsent,2present,| 8 | Sl 0 Bed 4 MS i aI 5 £55 | | | | | | | 4342 | present, 2absent,} 31 | 25 4 Bet a 3 0 eae eee ay fae | lias | | | : t | 453; | 1Lpresent,2absent,} 4 | 13 | 0 | 1 | 0 4 > F my |

3046. Hippodamia convergens 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 septempunetata and to a

_less degree to that in Adonia variegata. The varieties of these two Euro-

pean species have been well studied and named. This is not the case with the American H. convergens. Ihave 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 4, 1, 2, 3, 4, 5, 6. Typical.

Spots 4 +3-+1, 4+-5, 6, var. quilagesiqnate. Kirby.

Spots 4 +3-+1, 4+5, 6, with pronotal discal spots, var. puncticollis Casey.

Spots 4+3-+1, 2, 4-+-5, 6, H. septemmaculata var. continua.

Spots 4-+3-++1, 2, 4, 5, 6.

Spots 4 + 3,1, 2, 4-++5, 6, var. caseyi (new variety) <A. variegata var. ustulata Weise. => 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. defecta (new variety). Formula as in var. caseyi, but spot 1 small and1-+ 3 much less heavily pigmented.

Because of its interesting relation with var. caseyi, discussed later, this variety is given a name.

22

ES ee

HIPPODAMIA, 23

Spots 3 +3, 1, 4+.5, 6<> H. septemmaculata var. viadri Weise.

Spots 4 + 3, 1, 2, 4, 5, 6, <= H. septemmaculata var. paykulli Weise <> H. tredecim- punctata var. triloba Weise.

Spots 4 + 3, 1, 4, 5, 6, <> H. septemmaculata var. oblonga Weise.

Spots 3, 1, 2, 3, 4,5+6, <= H. tredecimpunctata var. c. nigrum; rare; California and Nevada.

Spots, 4, 1, 2, 3, 4, 5+-6, var. juncta Casey; taken in Sonoma County, California, and 1 specimen from Stony Lake, Michigan.

Spots 4, 1, 2, 3, 4+ 6, 5, rare; Pacific States.

Spots 4, +3+1, 2,4+5-+6, rare; Rocky Mountains.

Spots 3, 1, 2, 3, 44.5, 6-H. tredecimpunctata var. spissa Weise =H. septemmacu- lata var. aestiba; common in Western States.

Spots 3, 1, 4+-5, 6, pseudoglacialis (new variety); New Mexico and northward.

Spots 4 +3-+ e, extensa Mulsant, Western States.

Spots 4 +3-+ e, 2, 4, 5, 6. & Se Ss J “

Spots 4 +3-+e, 2,4+5, 6. b c d h

Spots 4 + 3, e, 2, 4, 5, 6.

Fie. 12—Variation of elytral pattern of the quindecimmaculata varieties. e=

mode; a, from Lake Superior; 6, California; ¢, St. Louis, Missouri; d, lowa

City, lowa; e, Southern Ilinois; i Kamiack Butte, W antieaeie h, Keeler, California; 7, "St. Louis, Missour'

Spots 4~++ 3, e, e’, 2, 4, 5, 6.

Spots 4 + 3+ e+’, 4, 5, 6, lecontei Mulsant.

Spots 4, 1, 2, 3+ gq, 4, 5, 6, quindecimmaculata Mulsant (fig. 12); Central States.

Spots 3,1, 2, 3, g, 4, 5, 6; very rare; Central States.

Spots 4, 1, 2,3+ q+’, 4, 5, 6, difficult to distinguish from 3+ g, but known to exist from a few specimens with reduced pigment.

Spots 4, 1, 2, 3+-q, 4, 5, 6, quindecimmaculata A Mulsant; very rare; Central States.

Spots 3, 1, 2, 3 + g,4+5, 6, quindecimmaculata D Mulsant; Missouri Valley; very rare.

Spots 3, 1, 2,4-+3+4q, 4, 5, 6, quindecimmaculata C Mulsant; Missouri Valley; very rare.

Spots 4 +3+-q, 4, 5; 1 specimen from Fairfield, Washington.

Spots 4, 1, 2-+q,3, 4, 5, 6; very rare; 1 specimen from Kamiack Butte, Washington, upon one side only.

Spots 3, 1,2+q+3-+ 3, 4, 5, 6; very rare; Missouri Valley.

Spots 4, 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=>= septemmaculata var. vorax Weise. Eastern as well as Western States.

Several spots lacking, convergens C Mulsant; Western States.

Mark 3 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 Fald. <> H. septem- maculata var. rubra Weise.

Cart

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. Ihave 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 upon H. 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.

a b a a A a

Fra. 13.—Variation of color-pattern of var. moesta, with two intergrades. 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 3 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 4 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 allof 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.

CbEh2 DebPD

Fig. 14.—Variation of pronotal pattern. a. b, d, e, from Dilley,

Oregon; ¢, Fairfield, Washington; f to j, Ber keley, California:

TABLE 4.—Elytral pattern in Hippodamia convergens.

| 5 Var. 12 spots) Var. quin- Glac- Spot-| Some | pres- ex- de- ialis Locality. No. “Por) spots ent;no) 4+5|34+3 ten. 1434+) Go t. 1+2 5+6 44546 SS- absent.) con- | é oe finenke! sa. mac- tern. ulata. Palo Alto, aL (K. P.ct. P.ct.| P.ct. | P.ct.| P.ct.| P.ct. P.ct. | P.ct. | P.ct.| Pct. P.ct. Pct. i) ere 033 > 1— 6.09 93.6 1— 0 | 0 0 0 0 I— 1l— ~ Kamietk “Butte WEAGI 3 eden ace 15,415 | 1— 15.3 591} 6.3 36.2 3.7 8.0 l- 1. #0 0 0 | Marsh se Pair- |— — | field, W fash...... ‘ _— 37 8.6 | 37.9 2.4 9.8 0 0 0 0 0 TABLE 5.—Elytral pattern in Hippodamia convergens. Per cent/Per cent ‘ > age he typical P pith cay, F ARS z er cent spots a spots é Locality. No spotless. sent; no! present; ya i con- no con- guent fluence. | fluence. . Chewelah, Wash............. 191 0 27.2 1— 72.2 | Mount Car leton, W: ash.. 359 0 — 0 99+ Spokane andC heney yw ash. . 107 0 58.8 | 5.6 33.5 | Fairfield, Wash... ........... 406 i= 33.7 | 8.6 50.2 | Steptoe Butte, Wash......... 573 0 81.3 0 18.6 | | Kamiack Butte, Wash....... 15,415 j— 40.7 | 10.7 48.4 | ' Goldendale, Wash........... 159 0 11.9 84.9 3.1 Portland, Oreg@.c)ic5 6 ii. ikcas 10 0 10 90 0 Dilley, Qreg :. 5: 255. Seen 895 94.9 4.2 1— 0 Berkeley, Cal..... 673 45.6 4.9 48.8 a Oakland, Cal., A. 632 9 2 1— 0 0 Oakland, Cal., B 7 31.5 0 68.4 0 Watsonville, ro aaa i? 68 51.4 0 48.5 0 Santa Clara Valley, Cal... =. 425 2.3 6.3 90.3 I— Mendocino County, Gal..:.. S 0 3.5 $6.4 0 Licking Fork, Cal............ 17 5.8 70.5 23.5 0 Redlands, Cal -.. ... 22.52. .28. 71 12.6 1.4 85.9 0 San Diego, € a a a tenes t 7a 0 2 0 Palo Alto, | Erne 1,033 1— 6.09 93.6 1.7 | San José,’ gs CTS TS 267 1.1 1.8 97.0 0. | Coolidge, N. Mex 75 0 14.6 32.0 53.8. | Tepexpam, | Mexico... ....... 201 4 17 7 1— Bartlesville, Okla............. 19 0 0 100 0 Stony Lake, Mich............ 20 0 0 95 0 Cold Spring Harbor, N.Y ... 203 0 1— 99+ 1—

EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

TABLE 6.— Variation of pronotum of Hippodamia convergens.

Marginal dash. Margin, ith- | Com- Elytral With| W | Con- Li lity. = No, out 7 | n plete lncom- ocality spots. dash.) dash. na Dh | Con- fluent land not ipecee Incom-| plete Suet fluent cepha-| con- janis plete | ceph- Tha’ |laterad ladand| fluent | SPA laterad alad dash. erad, P.ct.| Poet.| P.ct.| P. ct. P. et.| P. ct.| P.ct.| P. ct. | P. et Chewelah, Wash............ pos ab- 52 | 94 6 6 0 0 0 sent. / 0 DOsncatviiess +. - ae pome con- | 138 | 100 0 0 Oi: 10 86 0 14 uen 0 Mount Carlton,Wash.......|..... Do. 139] 99 | 1 1 o | @ 41 0 58 | 0 Spokane and Cheney, Wash.| Present. 6 | 100 0 0 ee ee 100 0 Ont Oo RS as ee et Cae ab- 68 | 94 6 3 0 0 97 0 0 | sent. | 0 DQ 56 caasighsandusaiesy Rome ope 92 8 3 | Oxf 63 5 29, | uen 0 nae Wash., 1907........ Present. 14 | 100 0 i oag) 0 100 0 0 0 NE BS Pea Sh 5 {8 Some ab- 94] 97 3 te ine 0 98 1 0 sent. 0 DOs. 05 cas vasieneaane oe Pome coe 148 | 100 0 A ee 0 ll : = “ uen Fairfield, Wash., 1907........]..... DOK, A 200 | 93 7 ae GS 0 89 0 Steptoe Butte, Wash........ oe ab-| 466] 98 2 2 | 0 0 89 0 9 : sen DO bizxs casos cee ee gel 107 | 96 4 Rte @ 0 86 0 12 : uent, | Kamiack Butte, Wash...... nt.. 118 | 9° 3 1 0 0 88 0 ll 0 Do. toe Some ab- |2,372| 9% | 4 4 i— ee 96 sen > A DOs. 3=gaeee age. 8 onan Poe snd 100, 7 3 0 0 0 9 se bet. 0 absent, Do... Raavetes geccn rian = on: 100, 1 0 0 0 oe |, 26 me re 4) uent. | Goldendale, Wash.......... Present. 154 100—| 1— 0 0 0 100— ss 1— a Portland, Oreg. 8 0 12 0 0 76 0 rem. 32 Dilley ,, OTOR.:. ac ce ck ta cnet 0 5 0 2 74 0 10 | 0 i aisle otra ea cent 1— 3 0 10 a ee 5 1-- Ber keley and Oakland, Cal ny : : . = | : ; . Pa eee | | gigs ied | 58 1— 0 0 16 | 9 4 7 Watioaels, Cal. a : : ° “ 0 2 > errr se : 7 Sante Chie Valley, Cal... 3 | 0 0 0 0 700 | #O 0 0 DO s..s8 se seeS eee | 3 0 0 0 97 0 0 3 100-55 .s een donc 3 0 0 Diy 2h a 0 Mendocino County, Cal..... Present 84 | 100 0 0 0 0 100 0 0 0 Licking Fork, Cal........... a ab- 12 | 100 0 0 | 0 0 100 0 0 0 sent. DO sts. shasteecet eae Present 4 100 0 OW, | 0 0 100 | 0 0 0 saci og Cal; . .. 2534) oeos eae “a «ewe = | i ; ; : “39 : ? ‘ = SORE See PPE Re ee re en | Coolidge, N. Mex.. Present. 24 | 100 0 ee ee | 0 |.100 | Q 0 0 DO Gicianie sPupie ae ann iaieae poe’ ab- ll | 73 27 Sy) 0 FH Gheod 0 9 sen | ree ed oes Some con-| 40 0 | 100 ee 0 10 0 0 0 uen Durango, Colo............... . ie ae 19|- 37 | 68 0 | 0 0 21 21 ul 47 Milford, Utah... 00.0020 20oh. es: 12| % | 7 0 0 0 92 8 0 0 Stony Lake, Mich |S. 2.25) Present... 19 | 100 0 0 | 0 0 1¢0 0 0 0 Cold Spring Harbor, N.Y...)..... BGs; 5... 202 | 100 0 aed 0 0 100 0 0 0 Tepexpam, Mexico.........|..... DO....553 200 100 0 5 ae 0 0 99 0 0 0 ji ORE ere Se 7,824 | 89 | 11 oe oe 1—| 80 | 2 8 8

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 } + 3, + 3 +1, 4 +5, and the absence of pronotal dashes and reduction of the light margin in the pronotum.

TSERS

Fie. 15.—Variation in color-pattern of head of Hippodamia convergens. c= mode; a4 toe, from California.

Subspecies of California, outside of the mountains and humid coast region, is characterized by a large percentage of spotless elytra and of pronota without dashes, and also of a reduction of the white margin. When the basal band is present it is of the variety extensa without the other spots.

Subspecies of Western Oregon, except the humid region, has elytra as in California, but the pronota with large dashes which sometimes open to the margin and with the light margins well developed.

Subspecies of the Great Plains shows in many specimens a reduction in the number and size of spots, occasionally entire absence of spots.

Subspecies of the Mississippi Valley, with Missouri as a center, comprise the only specimens of the variety quindecimmaculata and its forms (with the exception of two very rare specimens from the Pacific States).

Subspecies of the humid Pacific coast comprises the variety moesta. I have seen none of the lighter varieties in collections from the regions where the variety moesta is most abundant, but this does not establish their pres-

ence or absence.

One large lot taken from the top of Marsh Hill, Fairfield, Washington, contained 15,415 individuals of this species and its different varieties, and 759 individuals of H. spuria and its varieties. The various patterns of the H. convergens series were sorted and are represented in fig. 16. Those with no spots or with some spots absent but with no confluence are included in the 6,954 individuals of the normal pattern, as it is not feasible to rep- resent the combination of spots present in the figure. They are discussed on a later page.

The march toward greater pigmentation is orderly and determinate, but is not along one definite line of succession. The increase shows itself first generally in the confluence of 4 and 5 or 2 and 3. Whichever appears first, it is generally followed by the other. The coalescence of 1 and 3 is usually the next step. Further pigmentation enlarges these areas to a con- dition approaching that H. bowditchi. Only rarely does pigmentation go to the extent of confluence between 4 + 5 and 6 or between 4+ 5 and 2.

EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

odamia convergens in a hibernating mass of 15,415

arsh Hill, Fairfield, Washington.

“Ki

individuals from the top of

Fre. 16.—Variation in elytral pattern in Ai

HIPPODAMIA. 29

Quite a different line of development is occasioned by the appearance of spot e, the crescent-shaped mark just cephalad of 3 (fig. 16, nn, vv, etc.). This is the essential feature in the variety extensa, but may be seen with various patterns so far as the rest of the elytron is concerned. Ordinarily pigmentation not only makes spot e confluent with spot 3, but is so profuse as to hide spot e entirely (fig. 16, zz). For this reason I am uncertain whether spot e is invariably present in the true variety quinquesignata. I suspect it is a necessary character, without which 1 and 3 would fail to coalesce or coalesce only as in Nos. 421 and 427. Where +is confluent with 3, e is often present, but there are cases where it is clearly not present and others where its presence is doubtful. Where e is present and pig- mentation increases, confluence with 1 may result to form variety quinque- signata (fff). If 1 is absent, the pigment projects laterad to an acute angle (uu), the condition given the name extensa by Mulsant when other spots are absent.

Still another line of development is dependent on the appearance of another new spot which I call g, just laterad of spot 3 (ff). Itis very rare that where q is present it is not so widely confluent with 3 as to cause the impression that spot 3 is simply produced laterad (71). Such was my original conception. Another spot, g’, may carry the pigment still farther

laterad. It may even coalesce with 2, but the tendency is not a strong one, for the pigment frequently stops short of it, with a narrow line of the ground-color intervening.

Confluence of g with $ and of g and q’ with 1 is seen very rarely. Spots 4 and 5 are generally larger when q is present, and 4 is frequently squarish in outline. This makes it possible that we have in variety quindecim- maculata and its allied varieties a distinct species, although that view is not here adopted.

Spotless elytra are most frequent in the non-mountainous and non-humid sections of the Pacific States. Specimens from Onaga, Kansas, sent me by Mr. Crevecoeur, show the absence of several spots. Reduced pigment is also known from Pheenix, Arizona, and Black Hills, South Dakota. I suspect the specimens from the Black Hills were obtained from the plains around the Black Hills rather than from their higher portions, as we have the mountainous varieties also reported from the Black Hills. Apparently the Pacific Coast area of the variety obsoleta extends through Arizona to the Great Plains, where it exhibits itself more largely in transitional in- dividuals.

TABLE 7.—Number of spots in Hippodamia convergens.

| Locality. | . No. TH re 2 | 3 4 5 9 } Tae : Pict.) PP. ct. | Po ct. | P.ct. | Pct P.ct.| P.ct | Kamiack Butte, Wash...... 15,415 | 006; 0.20 1.08; 3.644: §.62/ 27.3! 59.2 | Palo Alto, Cal.............-. 1,033 9) 78). 1:46 VE ee, 22.3) 93.9 | Cold Spring Harbor, N. Y. 22; @ |/@ | 0 0 | 0 I—} 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 gq, of ontogenetic origin. The poly-

= 3 fo I:

Fra. 17.—Relative frequency of loss of the several spots in the genus Hippodamia. roken line, col- lection of Kellogg and Bell from Palo Alto, California; entire line, 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 novery 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. Herespot 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-

i

=a

= ris ; | 71 ; z o . Sy SIE paces A A ad ws Oe Ss ‘a Spots Sy iit Se SS Ge Sh ee ar ae a ee a Le Be ie cae Se 22 3 f 6 2 32 $4254 5368 62 6 4 B-S43° 25°65 3.4 6 $ . ma Ss e - r ) ee | ° 0 0=—673 kins 3583664 327 313 241241 168 157 73 72 41 R 2% 2191814109 8 TT 6 =F Having. ens 1 6.2 4-12 6 1 6.9 2 6 118-9 O13 6 61D Oe 1-8 D1 4 O—st

Fie. 18.—Correlation of loss of spots.

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 persistalone. 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 bowditchi, yet it is only rarely met with in cases of enlarged 4-++5 in the mountains. Itis more frequently found in Califor- nia in regions where the spotless variety is present.

TABLE 9.— Combinations of spots preys and absent in H. convergens from Fairfield, Wash.

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. Itis very rare, with- out intergrading conditions, and constitutes a distinct position of organic

stability. TABLE 8.—Spots lesidiie in Hippodamia convergens.

[Percentage of beetles in which each spot is lacking.]

| ye Locality. | No Spot 1. | Spot 2. | Spot3 | Spot 4 Spot 4. | PalovA lito, Cal .yctiewds ctecrns 1,033 3.5 | 0.9 2.7 3.5 3.4 | 3.9 | Kamiack Butte, Wash........ | 15,415 36.2 | revs 4.1 2.4 1.5) 9.3 Cold Spring Harbor, N. Y.... 0 | 0 0 0 er —1 | Sopenkests MGX1€0. 6.060 nc 200 110 | 7.0 | 7.0 12.0 13. 0) 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 toe. 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 1 eastern Wash- ington from that of western Oregon.

Present. Absent. | | P. ct. of Total. | l l | Spots. Tota], | Sum of | In In In | In | In | Sin- Sin- | _In In | In In In | _ totals | sixes. | fives. | fours ee eswpar? gly. gly. pairs. |trios.| fours. fives. apres | bret ee | | | | | | | | | 9,828 9,121 | 615 69 | 1 10° shee 1 3,583 1,258 548 | 156 32 10 | 5,587 36.2 14,779 9,121 | 4,166 | 1.131 | 283 69 9 3 52 196 278 97 25 10 636 | 4.1 14,233 9,121 | 3,871 975° | 219.) 44 -| 13 2 327 302 342 | 122 29 10 1,182 vr 15,191 9,121 | 4,191 1,247 | 512 | 107 13 5 80 49 59 19 10 | 224 1.5 15,048 | 9,121 | 4,190 | 1,265 | 387 | 80 5 4 8 62 | 174 | 86 27 10 | . 367 2.4 | 13,992 9,121 | 3,957 621 | 269 | 22 2 6 241 706 =| 202 , 144 30 10 | 1,423 9.3 Av 9,121 | 3,498.3 | 884.7; 280.5, 55.3 | 53]........ 60.7 | 442.3 280.5 110.7 267 10 | 9419 10.2 CORRELATION.

The correlation between the confluence of spots 4 and 5 with the con- fluence of spot 3 and the mark # 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 2 and 3 with relation of spots 4 and 5 in Hippodamia convergens at Fairfield, Washington.

| | 4 and 5. «a | Relation, 4% and 3. Gili i * P ne nfiu- | Transi- pa- | | ent. tion. rate, | Absent. Total. | Sere fe ag? oy } CMIARE. J... 2 oes a= cae aed *177 | 2 22 | 1 202 | Wem > 2 2. Ses eee q.j 0 6 | 1 8 | PORIEGED Se vc suc Ras <a cosane | 14 1 £9 | 4 118 | a ee eed o- 0 5 | tl 6 Dt Set Sop ae 192 | 3 132 7 334 | * Quind ; signata and caseyi. + Convergens. t Var. obsoleta,

TABLE 11.—Correlation of relation of spots 1 and 3 with relation of spots 4 and 5 in Hippodamia convergens at Fairfield, Washington.

| 4 and 5. Relation, 1 and 3. | | * . aa ransi- pa- * | Fused. | tion. | rate. | Absent. | Total. | | | | | | Oe ee Pe. Pe ee *64 | 0 0 | 0) 64 MVOMMTON. 3.653 «<3. dcs oc. 1} 0 0 | 0 | 1 SOPMPRIG Fs iccis ss aes nae +120 | 3 $59 | 0! 182 Absent.... 7 | 0 74 | 26 87 2 eee ee ee, ee 192 | 3 133 | 6 334 * Var. quinquesignata. T Var. caseyi. I Convergens. 2 Var. obsoleta.

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.

84 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

/\

(a) The distance between and confluence of /

spots 4 and 5 (in units of .625 millimeters.) | \ (b) Males, ‘aint ' (c) Females.

(d) All individuals.

| | | tetet Was 6 7 8 9 W 11 12 13 14 15 16 8 14 15 14 13 13 2 1016 213 7 7 6 18 31 36 27 20 20 8

3 2 0 2 a

| 5 4 062, 41 O= 5 10 1. 0= 9 42 2 0=

-~t

8 :

Fie. 19.—Variation ot the distance between and confluence of spots 4and 5 in Hippodamia convergens at Fairfield, 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.

———— SCO

HIPPODAMIA. 35

Fig. 20 shows the variation in the relation of the mark + to spot 3. The same conditions hold as before.

of the mark 4 and spot 3. (6) Males. 7 (¢) Females. 1 (d) All individuals. /

(a) The distance between and confluence (

' ‘ rT Py 1 | | A Sioa fd seg 4 = Re BS Pai (@) 1716 1514131211109 8 7654321012345 67 8 9 1011 12131415161718190 2 ®0053033136611052438300136147181416117902000 0 0-182 () 0112322555565438532022425 91881010159 § 610001 0170 (MW) 0165355690N11 7169 5129620838 50 7ISWAAWBGIT6 30001 0= 327

Fig. 20.—Variation of the distance between and confluence of the mark }4 and spot 3 in Hippodamia convergens at Fairfield, 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).

i | |

sy j \ 2 We i | ; <F <L--T~ (a) 3 2 21 219 181716 151413121109 S T6543 21L1LV0 Li 2S 456 TEIMNM @®M00013138362208322 42118316616 218010001113118501 4 OO—11 010114742427 40 £BEDWVB2Z1 23000 1120743101 0-19 @010145101046 2106 2 § 64139 4250101183818 5711 «5 O= 29

Fic. 2i—Variation of the distance between and confluence of the spots 1 and 3 in Hippodamia convergens at Fairfield, 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 ries specimens are at or near 5 units.

(a) Diameter of spot 1. (b) Males.

(c) Females. ‘ (d) AlLindividuals. j ‘

So - _ =

h NS :

| | 3 3 3

fa) 0 1 8 9 (b) 0 2 3 ce Reg fae (c) 0 3 3 4446561 (dq) 0 5 6 10 6 5 S36 *7 Fig. 22.—Variation in the diameter of sp Fairfield, W:

The variability in diameter of spot 3 is given in fig. 23. Here the typical species is at or near 4 units, and the two varieties just mentioned at or near 8units. The distance of the lateral edge of spot 3 from the suture is given in fig. 24. The lack of bimodality shows that the extension of pig- ment laterad in the variety caseyi is slight, the main extension being toward the suture in this case. The one individual at 27 units was a case of confluence with spot e.

Table 12 was constructed to test the correlation of these color-patterns with differences of size or shape. Wesee in them that the pronotum is broader in proportion where there is confluence than where there is not.

TABLE 12.—Comparison of ratio of breadth of pronotum to length of elytron in a lot of Hippodamia convergens from Fairfield, Washington.

| Variety. | Male. | Female.

| Without confluence of spots.. " 52.48 + 0.20 | 50.71 + 0.22 ot a eee ‘| 54.08 + 118 51.77 + .23 | Var. quinquesignata... .......: 54.06+ .39 | 52,41 + .38

i

was the elytral pattern shown in @) All individuals.

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 i\ pigment was obtained by expos- j \ ing the prepupa and pupa to the i| \ cold of a refrigerator (5° to 15° i | \ C.), a cellar (15° to 17° C.), and habia

the intermittent temperature of an ordinary room where the j | temperature dropped during the j i winter months from 12° to 18° C. i} kA by day to 5° C. at night. The i

greatest degree of pigmentation resulted in the last case. The result from one experiment, in which the prepupz of typical Hippodamia convergens were subjected to a freezing tempera- (a) Diameter of spot 3. | ture out of doors for two nights, (? Paks.

fig. 25, which is much like that of the western mountain beetles. The other cases give the prono- tum that is so frequently found in the western beetles, but elytra not corresponding to any known variety. Fig. 26 is a composite

drawing, no one beetle ShOWINE te) 0 12 3 $s 9 0 i is is the same high degree of pigmen- {3 > § 3 i Siemans 8 coe tation at every point. Some fea-“9® # 5 5 11 1923 0 127 1 O=118

tures of this modification pattern Fig. 23.—Variation of the diameter of spot 2 in Hippo- show a correspondence to some damia convergens at Fairfield, Washington.

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 confiuence 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

38 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

is clear that the end positions, that of no spot versus 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.

Fria. peter cn a of 2959,exposed to inter- mittent cold when a prepupa.

F1a@. 26.—A new pattern, produced by inter- mittent exposure to cold, which is not found in nature. From Hippodamia convergens.

: , | Se (a) 10 11 12 13 14 15 16 17 18 19 20 21 (> 0 1 2.4 12 15 8% 8 2 8 8 1 (c) Or 0 0 FF BeBe Sew BS. 4.0 (q) 0 1-2 114 4-72 7 40 11 4 «#1

28

0=139 0=150 0 0 4 0=239

Fre. 24.—Variation of distance from the suture through spot 3 in Hippoda- mia convergens at Fairfield, Washington.

(a) Distance from suture through spot 3. (b) Males. (c) Females. (ad) All individuals.

5) 26 0 0 0 0 0

coos coow

Owing to the small number of extended pedigrees, the extent of domi- nance and segregation in these beetles must 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 acharacteristic 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 the12 females of this kind which had mated in nature, we find that not infrequently a few

ee

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 3, spot 6 present 5 + 3, spot 4 present 9+ 4, spot 3 present 8+ 4, spot 2 present 11 + 3, spot 5 present 16+ 3. 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 +4and 6+ 3, 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. |

Mother 4 A. | Mother 4 db. Pricg b. Mother zz. Mother a a. } AA 100 100 | 100 | 4 PEE ace in at pine 5's <5 100 i 75 | 100 Or 75 AAorAb...... 100 1o0or75 | 100 or 75 es trreccucent 100 ; nO 100 or 50 3 Re ee 100 | 5SO0to 100 | 100to SD I ete en coos a oh ; 100 Rr 50 with 50 inter. | . ee re | 100 inter. i WHS hotissati de 100 ot ea 0 Geass einen sees j i Bt | A A ora A....| dor ee ieee - | 0,50, or 100

Spots 1, 2, and 6 are often solely absent in regions where spotlessness prevails, so that their individual heredity becomes of interest. Spot 6 was not lacking in any of the offspring of No. 2282, which was so character- ized, but of the similar No. 232 2 the single offspring had the spot reduced. In No. 455, where 6 was transitional, it was normal in all the offspring.

40 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

The presence of this spot is, then, either dominant or preponderant over its absence.

Spot 2 was also either dominant or preponderant, for where the mates were opposed in this character, presence of the spot in the offspring always prevailed. Inheritance of spot 1 was studied in material from Fairfield, — Washington, where it is frequently lacking. Its presence is less strongly inherited than that of the other spots, being found in less than half of the progeny in the case of presence < absence. I believe that this result has a connection with the fact of its frequent absence in that locality. This is in contrast to the inheritance of spots 2 and 6.

The size of spots is inheritable, as seen in fig. 27, where the unusual proportion in size between 4 and 5 is inherited in some degree. This is conspicuously true of the size of spot 1, where we have, besides the normal center of variation, a second, made up of varieties quinquesignata and casey?, having the spot large. The shape of spots is also inheritable; thus in fig. 27 the influence of 4+ 6 in one of the parents shows itself in the progeny, though they lack that confluence.

| oi?

Fig. 27.—Elytral pattern of offspring of 2599.

The confluence of 4 and 5 is regularly inherited in the heavily pig- mented individuals, especially in the variety quinquesignata. In more lightly pigmented individuals its inheritance is more irregular. It acts as neither a simple dominant nor recessive, yet segregation is evident. The suggestion that a greater degree of pigment is always dominant over a lesser degree does not receive support.

The confluence of 4and 6 has neither dominance nor trimorphic heredity. It is segregate,* however, and apparently subponderant. The confluence of 4 and 3 is inheritable in various degrees, dependent largely upon the degree of itsdevelopment. There is segregation with some intermediates, yet the suppositions of dominance or trimorphic inheritance are not tenable. The presence of the extensa character is difficult of analysis, for its presence may be cloaked with excessive pigment. Yet, bearing this in mind, we see that while there is segregation with some intermediates, neither domi- nance nor trimorphic heredity are found.

* The expression ‘‘alternative heredity’’ is a misnomer in cases where the heterozygotes are intermediate and I therefore prefer the term ‘‘segregate heredity.’’ This is divided into amphimorphic heredity, where the two parental conditions are represented in F., and trimorphic, where, in addition, we have a third intermediate condition in F,.

ee

HIPPODAMIA. 41

All cases of the confluence of 1 and 3 are not inherited alike. In matings of No. 4212 and No. 427 2, the confluence is the result of the over- growth of land3. There is no mutual projection of the spots and they are not confluent when the spots first appear in ontogeny. Such cases of confluence are not inherited in kind, because, being only extreme fluc- tuants, the offspring do not attain their magnitude, and therefore no con- fluence results. In the cases of true confluence, again, there are two kinds which are inherited differently. In one kind, typified by No. 351 2, spot 1 is small, so that the basal band tapers toit without an enlargement. This is rather unstable and its inheritance has many of the irregularities

TABLE 14.—Pedigree table of a race of the variety quinquesignata.

? x “lecontei” ? x “quinquesignata” 7 ee — a 1 1 rT 3 9 1 7 a convergens (4+5) “lecontei” quinquesignata ($+3+1) “lecontei” 7 == =

Scr x x

| ) | |

20 21 15 5

quinquesignata quinquesignata quinquesignata quinquesignata

x

! quinquesignata

of the preceding characteristics. In cases of the confluence of 1 and 3, where there is heavy pigmentation with spot 1, showing as a conspicuous enlargement, the variety quinquesignata (3502, 3822, 3902, and 42292), the inheritance is unique, showing a high degree of stability and segrega- tion. The results (table 15) are not inconsistent with the hypothesis of dominance, but in 3822 and 439°, where we have a transition to con- fluence of 1 and 3, there are nonconformable results, so that even if there be dominance here it is not firmly established.

Explanation of Table 15.

The letters in the columns headed “ Father” and “ Offspring of” indicate that the individual was unknown, but must have been from the locality indicated, as follows: B., Berkeley, Cal.; C., Chewelah, ash.; C.S. H., Cold Spring Harbor, N. Y.; D., Dilley, Oreg.; F., Fairfield, Wash.; O. B., Oyster Bay, N. Y.; P. A., Palo Alto, Cal; R., Rediands, Cal.” The prevailing patterns at these localities are given in preceding tables. : _

F or M in the remaining columns places the father and mother, respectively, in the category named at the head of the column. In other respects it was a typical H. convergens. jf indicates that the sire was used with a female that was probably already fecundated and hence was improbably the father. These series were included in order that their characteristics might be preserved for subsequent families; but these were rarely obtained. : : —

The numbers under “ Father” and “ Mother” are designations of the parents; in the remaining columns they indicate the number of offspring in the several categories. 5

* Indicates that information is lacking as to the character of the pronotal markings.

(1) Pronotum of M modified by cold. Offspring emerged in autumn.

(2) Spots 2 and 3 are larger in progeny than in parent, pe are smaller than normal.

(3) The light basal mark very prominent in M, normal in progeny. _

4) In M, 4is very near to5. Of3 others subjected to cold 1 had 4 tr. 5. {3} In M, 1+ 8 became slightly confluent after several days.

42

EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

TABLE 15.—Heredity in Hippodamia convergens.

( See bottom of p. 41.)

Father.

Mother.

Offspring of—

No.

Ss

sands

land 38.

4 and 5.

4and 6.

Ex-

Spot 1.

Tran- sition. +

xe) Beonnn imag =

3 ae Q Z =

2)

a o> Fy

hoe >ptt >

CQ Bessy ¢ Oo

Q sp DNy

cS met

St +OO

Sonn TM ee

oo

ala Mal alt Prite

bf

QO

nee Q 20 BORRyme Wigs = a om

Quan =

ole)

EF

tess

wo ~1 o

CSH

oy

373

Q

eSeh sarees

8

= s

Q

Q fe) mR DR 2eStane

- @ao Siqgy CWWeoosy yy roo™ tbh ae

2 wD x

"eeseen

235

RSaegyyayS

£ Fa at

_ —_

So mood

Seok

Noe Se Sewanee

wo

Bosom exkses

10

—~

—_ .

. 3 i + ©

ase |

La dpe |

ey ee

2 FM

SSSEE_: Sh CC

>be] ss as

Tran- sition.

ee. ae Se

Tran- sition.

a

Tran- sition.

tensa.

Tran-

sition.

“Ab- sent.

~

Ba: mm! co: coro: comoono: aS . : ° . ‘

PL ene ae

mdr tt:

ra Lite

| 3 MS a> | ave . ees be

wei THEE

OO ee ;

> OO! ROM! COLertoco: : :

et) tee ee er

eRe wel mes :

Rake sReeeee: 22:5:

>| = oo

: oO

ty ble) Bean sione! ERE: o = GO

Als

fil M19

ites

ad f : r ; cp 3 Pe : as Fe e6 a 7 * seg | * & * *x5 Z os oo +h bp - 4 z - 2 S Ey BP 0! Ea aS -) a ¥y z maf = ~~ 3 33 oe B é awe Rw ae EO OY WBE ES xX tng* Es ¥ went —* Exe ~ ot H*_ y XESS _xHKE _% ¥y Z ex x88 Sx B25 CReC Fas 2 RGSS SR Ait Scale ey Bt yo Oy hie bt Te ad $2 tm ee ae Shittttrirr tS Gas: wba ie aS SS a asi Se NGL ee tht a ees S Ost t Wee he har ieee pl RRS hae serae Saal = Bee as fe ERE EE " rare Pidy he ot tA e ie ee vd) eae Pittpeiie titittiawe Mim iiiisiiiiii: & a . ie ree eee.” Od Sr Lal wte aie. oe ive, C9 wee vis le CE Sh ww TE . . | eer Are)! ae See | Cite W ad” UE Ce om ee = FP) 0 1S wae) TOP we 8) ee Cr | . * . aw rs | & | Piiifibliitiies yg iti: ee ee eee | Se et Ces PO cee Be ae ee ey FE a | PELi ep ph eye bb ee og, “ ae ee ie cm 5 Sti Sgt $itr dS : E | Comme ks ot kee Sy ete Se eg ers : i PiiiMbintighh, Giitiediciiwetieniriity : 1 Dm ae OMDR Mig | am

| Ole ar | ! . ‘ (fe Lies hoc) Pilitl Pease Sosa i fae Aaa |

HIPPODAMIA.

ity in Hippodamia convergens—Continued.

ered:

TABLE 15.—H

3 ot 3 fi fia ay R eo Eee bebe We te fe tts mime; ferns | . a wy . eo. . et : vay Da e 8 8 ek Ft 8D mt oh] eheodawt: + ky» eos se Cee se eae oe epee 6 ~ soe + ommd ee ee, en ee ee 3 SAC SAE blood. pk OF ane meee LS Pe RI nt ee \ i ee 64 Oe EE RC EREN SET ERE | eb es Fe Oe ee Bea) ed . 0 oe * Are Se Fr A ae ee ae See es A oot Sie eles met ee : Pit bid pibigsiiisssie: a ea ge SS OR eles ion] M'QcorDayin + + wi< 8 ee Oe aviag ewe ues . 6 ~ . en | . - a ts see Sis Basses i: iit biti ibimipissssss i PiGiLilliWiii des deiii bigiiiiiiiiiiie Ce Re ee ec eh et a . ‘ Rr ae eos . by ae ee 8 EE dracotol™: johe! EB ats hat yids gree a ee ae a a i Sis) saan MS Rem et tee ae eee S B Sy Pea eT Nek Bo nie a See See Mae he BMS TNS Gee ae ing AMEE Mary gL REO ROS. ely eee MIE Cae o.'¢ a ar a-se 56 » 1 ee * . . ~~, © . . we a SEE TE Sen opie ee et — Rat hal” ee oP a of PEGE E Stops Aaaees : : eR Be hae Boge as Bi, H wa Se So ee } Swe “ei Re OF ‘A . a, Le . A eh + ME 6 . PAR re Leh eer eget: ee! ee aS ee ne a Tt

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 1492, 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 tolgive us continuous variation. Itis 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 casey: of full pigmentation and eastern specimens of H.convergens. But there was interfertility between specimens with the same confluence in a lesser degree and having spot 1 small (the variety defecta) 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, d) and 5 cotypes (fig. 28, 5. c, e, 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 muchenlarged. 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.

> Sesssse

Fre. 28.—Variation of color-pattern of Hippodamia bowditchi. d— mode; 6 to g, from St. Maries, Idahe; A, from Bitterroot Mountains: a, a specimen of H. convergens var. quinquesignata collected with them.

With the exception of a very slight convergent dash on one side of one specimen (fig. 28, a) the pronotum is spotless. The black pattern reaches the cephalic border of the pronotum on each side and in some cases extends along it for some distance (fig. 28, d). The black pattern always shows a lateral extension which generally reaches the border and in some cases extends to the caudo-lateral corner (fig. 28, 5). The variation of the pro- notum is paralleled by individuals of H. convergens.

The ground-color is yellow without a trace of reddish color, in strong contrast with that of the accompanying specimens of H. convergens. The peculiar ground-color suggests the possibility that these beetles are only larvee of H. convergens modified by some unusual food. Suchan hypothesis must be rejected for the present for the more likely one of a specific dif- ference. Unfortunately, I have been unable to attempt hybridization be- tween this new species and H. convergens.

46 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

A specimen from Kalispell, Montana, and one from between Hope and Okanogan, British Columbia, in the collection of Mr. F. K. Bowditch, belong here. A specimen in the Museum of Comparative Zoology from the Bitterroot Mountains has the characteristics of this species and, in addi- tion, a slight confluence of 4-++ 5 and 6 which is subordinate to the extreme confluence of 4+ 5. Further collections in these mountains adjoining the type locality might give us in this species a still further development in this direction.

3047. Hippodamia spuria Leconte.

Distribution: Plateau and Pacific States and British Columbia.

This is a variable species (fig. 29), but the forms show little geograph- ical differentiation. H. sinuata and cockerelli are specifically distinct derivatives, so far as our present knowledge goes.

Typical pattern (fig. 29, a): Suture not black, except at basal quarter. Spots 1, 4, 5, 6. 256 individuals in 759 —34 per cent, at Fairfield, Washington. Var. lineata (n. var.). The scutellar spot is replaced by a black suture only slightly wider at one-quarter. Intergrades occur, but they are less common than the variety. This char- acter is found in conjunction with most of the varieties of elytral pattern; 3 in 759 with 2 transitional, at Fairfield, Washington. Var. a, Spots 1, 4+5, 6. 447 in 759 —59 per cent, at Fairfield, Washington (fig. 29, h). Var. B. Spots 1,5-+4-+6. 37 in 759 =5 per cent, at Fairfield, Washington (fig. 29, k). Var. complex Casey. 1+4-+5-+6. Scutellar mark short. Not at Fairfield, Washington; Vancouver Island (fig. 29, 7). Var. crotchi Casey. 1+4-++5, 6. Scutellar mark short. 12 in 759=2 per sole at Fairfield, Washing- ton (fig. 29, z). Var. y. Spots 1,4-+5-+6. Rare, Idaho. Nowhere established (fig. 29, ~). Var. «. Spots 1 absent, 4, 5, and 6 reduced. 2 in 759 at Fairfield, Washington; Colorado (fig. 29, q).

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

Fie. 29.—Variation of elytral pattern in Hippodamia spuria.

DPRDPR =?

Fie. 30.—Variation of pronotal pattern in Hippodamia spuria. Mode =c tos.

Fie6,. 3l.—Variation of head-pattern of Hippodamia spuria from Dilley, Oregon. 6b = mode. only 5in 759 at Fairfield. On theother 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. 3682 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. convergens (3752), it is probable that its lack in the offspring here is the result of recessive- ness or subponderance.

TABLE 16.—Heredity in Hippodamia spuria. (For description of form of table and abbreviations see table 15, pp. 42, 43.]

Father. | Mother. | | Sutural pigment. land 5. 4 and 6. 5 and 6. | | No. | / Of Off- | Inter- . Tran- ‘ ; Sepa-| Ap- | Tran- Sepa- - Se Tran- No. paring No. - ing Short.) ne Long} hate, proach sition. + |arate. Fat + | rate. |sition. F F [sc] F | 48 |Mir| 1 |... is Wie... 8 Rests Titec kibes Mis} 0 | 4 D D |27:| D ef BS eras pe re ee Se ae eat ey PS M7 Gai Cee 262 3 a7 262 ¢| 227 | 5 |RM5)......|..... 2 We ae pee eee meee aes PMS). Bot... = F F |400¢| F 18 | MSL ose Waxstea 621 e Pees [setts [ovens fM1l5, 1 2 3 és | | eid F |8889]° @ 9-98 |f8028)......) 0 9 |fMir| 1 a ae FM2| 28 |...... 46s | 383 (461c| 383 | 3 |FM8]......] 0... canes [SR ot clit MR ctieces ita es FM38| 38 |...... eed 4894 | 883 [4809] 40 Bl BP eee cor to ore eee a aa | Bank (eae Gal don By 9 ie eee Se ; wig, F |80¢| Fa) 44 | £10) 6 | MO} 44.)¢ ssl. fM11| 7 | 26 i LA g. | | 4373 | 380 | 487¢| 380° | 4 |FM4/......]..... ot Re PDs, Sela. Pee FM4/0 4 |..u.. 4423 | 380 | 4429] 380] 15 | FO | MS} 2 16 |... hee. | ee FMS5!.3 | 7 > Psat. 4453 | 380 445¢| 380 | 9 4 |FMI1| 4 ess. ea eae Pa We Se... 4324 | 380 4329] 380 | 12 6 0 SRM BS A Seen Ree ee we ke Lee eS 4332 | 380 433¢] 380 | 5 2 1 MOMS BY csc Rencsn | eee | CEE (OMS) & [i0...

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 4and 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.

Forma. Spots $+1, 64+4+5, Pinnacle Peak, Washington.

Sosee 0?

Fic. 32.—Variation in color-pattern in H. oregonensis. c—mode; b, from Oregon; ¢ to e, from Pinnacle Peak, Washington. Fig. 33.—Variation in color-pattern of H.cockerelli. a, from Colorado; 6, from Yellowstone Park.

3048a. Hippoidamia cockerelli, new species.

Distribution: Colorado to Wyoming.

This species resembles H. spuria, except in color-pattern (see fig. 33). This is remarkable in possessing vittz 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. spwria, 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. 8. 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 4+-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.

> SOD

Fra. 34.—Color-pattern of H. dispar from Colorado.

Fie. 35.—Variation of elytral pattern in H. sinwata. b=mode; 6 to e, from Alameda, California; a, variety crotchi.

a

$049. 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 vittze complete, is, however, distinct from the other beetles often called H. sinwata, which have the vittze 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 Linnzus.

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). Thisis 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 (ef. fig. 37). Elytral spots small.

Var. signata Faldermann. Black pronotal pattern narrow. Elytra spotless. One from Washington, but with mark 3.

a b c d

Fig. 36.—Variation in elytral pattern in H. tredecimpunctaia. b= mode; a, from Washington; b, c, from Montana; d, from California; e, from Massachusetts; 7, from Oregon.

ad Pee EE

Fie. 37.—Variation of the pronotal pattern of H, tred:cimpune- tata. d=mode; a to e, from Stony Lake, Michigan. / toJ, from Pacific States.

SESE

Fie, 38—Variation of color of head in H. tredecimpunctata. b= mode; ato d, from Stony Lake, Michigan; e, from Cali- fornia.

Var. laeta Weise. Lacks spot 2 and 4.

Var. undecimmaculata Harrer. Lacks spot 2.

Var. scutellaris Weise =D Mulsant. Lacks mark 34.

Var. triloba Weise = F Mulsant. Spots 3+4.

Var. spissa Weise = G Mulsant. Spots 4+5.

Var. marchica Weise. Spots 1+ 2,4+ 5,6. Not seen in America.

Var. contorta Weise. Spots 1, $+3, 445, 6.

Var. c-nigrum Weise = H Mulsant. Spots 4+5-+6 in form of letter ec.

Var. gyllenhali Weise =I Mulsant. Spots 3, 1, 2, 3, 4+5+6 in the form of alarge angu- lar patch. Not seen in America.

Var. sellata Weise. Spots 4+3+2+5, 1, 6. Pattern saddle-shaped. Not seen in America.

Var. borealis Weise. Elytra black with one or two reddish spots. Not seen in America.

52 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

Var. A Mulsant. The sublateral spot of pronotum confluent with the central pattern (see fig. 37). In a lot of 111 from Stony Lake, Michigan, 80 per cent were of this variety, 10 per cent with a sublateral spot separate, and 10 per cent transitional.

Var. B Mulsant. Several of the black spots absent. One from Garland, Colorado.

Var. E Mulsant. The mark 3 prolonged to level of spot 3. Rare.

Subspecies of Oregon to California: Specimens from this region show spot 3+, spot 2 absent, and 4 small. The pronotal pattern has an ad- ditional lateral process cephalad of the usual one.

3053. Hippodamia americana Crotch.

Distribution: Lake Superior and Hudson Bay.

The type specimen in the Leconte collection makes it clear that this abused name belongs to a species closely allied to H. tredecimpunctata rather than to specimens of a variety of H. spuria. It is essentially a black-sutured, usually vittate derivative of H. tredecim- punctata (fig. 39). The specific rank must be retained for the present because of the lack of a large series. The great variation in the few specimens known makes it possi- ble that larger numbers from Lake Supe- rior might reduce it to a variety of H. tre- “i "americana, a. trom Tudou Bars decimpunctata. The pronotum has the pig- ™ sa le a ment so extended that it generally conceals the characteristic lateral pro- jection of H. tredecimpunctata.

a b c d

3051. Hippodamia parenthesis Say.

Distribution: United States, except the southeastern States.

Fitch has named a long series of forms, some of which are scarcely worthy of recognition (see fig. 40). Being based on one character, two or more of these forms are often embodied in one specimen. There is more intergradation in this species than in the others of Hippodamia.

Form tridentifrons Fitch. | Three white spots on head confluent. Common. (See fig. 41.) Form permacrifrons Fitch. Three spots on head, separate and reduced. Form triangularis Fitch. White spot on pronotum, triangular instead of square. Rare. (See fig. 42.) Form lituricollis Fitch.

Black pattern of pronotum much reduced, not touching caudal edge of pronotum and “‘having a large black dot on each side confluent with it by a neck only one- half as thick as the dot.’’ This form is interesting as showing components of the pattern similar to those in H. tredecimpunctata.

Form albomaculata Fitch. White basal spots adjoining scutellar marks.

HIPPODAMIA. 53

Form linearis Fitch. The scutellar mark linear, not dilated. Form approximata Fitch. The humeral spots touch the base of the elytra. Form disco-punctata Fitch. Spot 5 thick and round instead of oblong. Form tridens Fitch. : Spot 5 separate from 4+6. In many localities this exceeds the type in abundance. At Stony Lake, Oceana County, Michigan, 29 per cent of 652 have spot 4 separate, 3 per cent transition, and 68 per cent confluent in some degree. Form confluenta Fitch. Spots 544+6. Strongly confluent, only a narrow cleft dividing the masal ends of 4 and 6. Form connata Fitch. Spots, 1, 4+5+6 making one large patch. Form insulata Fitch. Spots, 1, 44+5-+6 inclosing a small light spot. Form nimia Fitch. Spots 1, 3, 445+6. Fitch reports 3 in 40 cases. At Stony Lake, Michigan, there

were 16 per cent in 657. e d e

Mulsant gives 2 varie- ties characterized by loss & of some spots, but I at- tribute these to Hippoda- “| @% mia apicalis, where the oS a 6 f a part of its range is

tendency to reduction in se CPPes a@ b e d .

e

Subspecies of the East- ern States: Characterized

by a large percentage of /[ forms confluenta, insu- Ig he he”) OD a lata, and connata.

a 5 c d é

Subspecies of Oregon: )

The black pattern upon the pronotum, instead of being only slightly wavy teat é i F f

cephalad, has the mesal

Fie. 49.—Variation in elytral pattern of H. parenthesis. 6 = mode;

element strongly project- _—3! fom Stony Lake, Michigan.

. Fre. 41.—Variation of color-patt f head - apicali

ing on account of the re- foe Colorado. Variation « or tandlawretig 53 hep; ‘H. ape duction 0 f pi ent be- wnat a ee e, d, from Stony Lake, Michigan; e, from

tween it and the lateral Fre. 42.—Variation of pronotal pattern of H.parenthesis. ¢ = mode; dlesnont atoe, Yy, Oregon; / toj, from Stony Lake, Michigan.

Although this species is so widespread, it seems less closely allied to the European members of its subgenus than the much more locally distributed species H. apicalis and H. lengi.

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 +3, 1+2, 44+5+6+ a. (Fig. 48, e.) Var. Spots 4+3, 142, 64+4+5+6-+-a inclosing alight spot. Arizona. (Fig. 48, h.) Var. Spots 4+3, 1+2+4+5+6+ a---H. arctica Schneider. Nevada and Utah. (Fig. 43, 9.) Var. Spots 4+3, 1, 5, 44+6+a. Idaho. (Fig. 48, 7.) Var. Spots 4+3, 1, 4, 5, 6+a. Colorado and Wyoming. (Fig. 43, c.) Var. Spots +3, 1, 4, 6+a. Colorado and Wyoming. Var. Spots $+3, 5, 4+6, a. Colorado and Wyoming. Var. Spots 4+3, 1, 4, 6, a. Colorado and Wyoming. Var. Spots $+3, 1, 4, a. Colorado and Wyoming. Var. Spots $+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.

S t b c id e f g A. Fiq. 43.—Variation of elytral pattern in H. apicalis. e = mode; a, from California;

b, d, from Colorado; ¢ from Wyoming; e,/, g, from Nevada; h, from$Arizona; i, 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 falcigera Crotch.

Distribution: Hudson Bay, Great Slave Lake. es

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 vittze 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, b.

Var. C. As in fig. 44,¢. The pronotal spots confluent.

5° PD

Fia. 44.—a to ¢, variation of color-pattern of NW. venusta. c, from Louisiana; d, e, NV. notulata.

3055 b. Neoharmonia notulata Mulsant (fig. 44).

Distribution: Louisiana.

Unfortunately Iam unable to throw light on the interesting question of the nature of this melanic derivative of the previous species. Itis 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 vittze (fig. 45).

The elytral pattern (fig. 46) is much reduced, revealing in some speci- mens 4 separate spots instead of 2 in the intermediate row and 4 separate

Fra, 45.—Variation in pronotal pattern of N. ampla. All from Brownsville, Texas,

Fria. 46—Variation in elytral pattern in NV. ampla. 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

Genus COCCINELLA Linnzus. The species of this genus recognized in the

TRIFASCIATA 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 @ vornicors 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 @ ) sutures 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, 7). 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, 7). Var. juliana Mulsant. Intermediate and caudal bands missing (fig. 49, f). 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.) Fi. 47.—The species of Coccinella Form F. Intermediate band only present. Oregon in the United States. (fig. 49, 0). 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 d).,

PROLONGATA

=>

NOVEM-NOTATA

JOHNSONI

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. SUDvEra~ .....5sisicc Ret ete ds swish Doge areata ede rhee aa eee 1 Var, €Ugenit..........:sdeikietisd: dda cedugiens beac bees) eee apes ae < Var. eugenii transition to var. JULIANG..........cssrceeecsesereccners 8 Var SFG 26,0800 s05s08s bose >onupeehiakaeed nates arches sateeedine te eee 78 Var.“ ytiana transition to Forma. Clr..sccvsscxconnesserrsesdenncances 6 PORTED, Jive ssinesccssedennansecspe ndvroscasateteestomner sa taateershoenstder tere 5 Porm. A,...2000cssncessseeeststecess <cpkitaes’sedanses (eae scttedaaenanad 1 Var. juliana transition to var. perplead...........cccsesereeeenesens 2 TY DiGREis «» 20c....0.0s0cdaveahvsisaavenntalemanh ce <xess0tceeeeeieie sess sienna 2

Large numbers show that the variety ewgenii is not as markedly set off from the variety juliana as has been stated. This is indicated by the intergrades noted above.

>>P DP

Fra. 48.—Variation in pronotal pattern in C. perplexa. b=mode,

The relationship to C. novemnotata suggested by Mulsant is not close, for the larve 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 and F.

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 sig 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). h

Var. spots H+. Var. spots [+-A.

Var. spots 3+. Seen on one side only 9 of 1 specimen from Oyster Bay, Long Island.

Var. spots L+H-+1. .

Var. spots L+H-+I+.A, confluenta , ‘ * Fitch. |

Var. divisicollis Fitch. Black pattern of pronotum touching the cephalic ae

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). ec \ ;

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 b \ [i k further study.

Form A. Some spots lacking.

Var. franciscana Mulsant. Elytra @ without spots. Fie. 49.—Variation of elytral pattern in C. perplera.

Subspecies of the Plateau States: Variety degener prevails. Subspecies of 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 prepupze and pupz 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 ina ratio of 4 per cent, and each of the other varieties was found in less than 0.1 : a per cent, the mates of females from nature

Fg. 50.—Variation in pronotal Were probably normal. The progeny of such pei pamode Ue movemno- 4 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 4402, and with the presence of such intergrades in nature. In general, the conclusion is

TABLE 17.—Heredity in Coccinella novemnotata.

{ Father. Mother. | h tr. 1. h-+l | No Nor. h+l htr.l 4 ’ h+l+m No. | Elytral a Elytral Atr.m htr.m * |pattern. 0: pattern. ae inar: in ee 278 2 Kiétin! 20 4 0 0 0 0 8 $ t+ 1K trae 440 2 | h+l+m 55 23 3 0 0 19 4 8) tt | 4052] nor 46 39 5 0 0 0 0 *449 2) m +a '*4499| mtra 14 6 0 0 0 0 0 -,|hap.l - -- 465 9 hap.m 24 16 2 5 1 0 1) Totes so es. FE 159 88 10 5 1 19 12 | Father. Mother. Atl h+l+m rh+l Ati+ No. htr.m |mtr. a) m+a No Elytral No Elytral yp mtr. a CS mee m+a * |pattern. * |pattern. a _ 278° |h+i+m 20 5 0 3 0 0 0 440 ¢ Atl = Vien m 440 2 |hA+l+-m 55 3 0 1 1 1 0 4052) A+ ; \+—| trem 405¢| nor. 46 0 0 0 0 0 2 449 4 m-+a \*4499) m tr. a 14 0 1 0 0 2 5 | ase ole HE | _ cae 465 2 | 7, ap. m 24 0 0 0 0 0 0 PE ood Dedaaaheies 169 | 8 eg ae 1 3 7

*Progeny of 405.

justified that there are here a number of positions of organic stability, namely, h+1, h+1+m, m+a, h+1+m+a; h+m was never seen without other confluence either in these experiments or wild. In progeny from 465 2 (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 finely 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 tomy knowledge. A large series is necessary to fix its status more definitely.

ULBD REDE

F1i4, 51.—Variation of elytral pattern of C. novemnotata in tne Western States. /—mode.

: a B. %, .

g. é

‘DODD ANeRRER EDEES it

wi

Fi4. 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.

i

pe.

BE MNNN NSD SD

F1i4. 53.—Offspring of 465 ¢.

DOSED EE

P14. 54.—Variation of elytral pattern of C. transversogutiata. a, from Spokane, Washington; b toe, from Kamloops, British Columbia; /, from Stony Lake, Michigan; g, from Wollaston, 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 novemnotata 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. 4-+ median band. Califor- nia and New Hampshire. (Fig. 58, d.) Var. postica n. var. Median band + apical band. a b ¢

California. (Fig. 58, c.) Var. confluenta n. var. %4-+ median band + api- clog tyrtuhtedeecner eo a ge A cee ot cal band. California. (Fig. 58, e).

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.

JD tp ppp

Fie. 5.— Variation of elytral pattern of C. californica. b=mode. Fig. 57.—Variation in pronotal pattern in C. californica. b=mode; d, variety melanocollis.

[22 )e De

Fie. 538.—Variation in elytral pattern in C. monticola. b=mode. Fic. 38.—Variation of pronotal pattern in @ monticola, b—mode.

Var. alutacea Casey. Pigment of median band so reduced as to result in complete or partial isolation of alateral 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, 5.)

Var. biguttata. Pigment much reduced. Lateral and apicgl 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.

PPP PPE

Fria. 60.—Variation of elytral patternin C. monticola var. alutacea and biguttata. F14. 61.—Pronotal pattern of C. monticola var. alutacea. a, asingle specimen.

3061 b. Coccinella suturalis Casey.

Mstribution: 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.

= c d

Fa. 62.—Variation in color-pattern in C. prolongata. a, b, c, from Cheney, Wash.; d, 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 Grotch.

The color-pattern and its lines of variation in this genus are markedly different from Coccinella, of which Leng makesit a subgenus. The form, agility, and the difference of the food of the larve (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 Linnzus.

Distribution: Cuba, Bahamas, Florida to California. Var. immaculata 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.

| dod eevee

Fie. 63.—Variation in pronotal pattern of Cycloneda munda male. a, 6, from Cold Spring Harbor, New York; ec. d, from Stony Lake, Michigan.

Fig. 64.—Variation in pronotal pattern in Cycloneda munda female.

3064 a. Cycloneda munda Say.

Distribution: Northeastern and North Central States.

The males havea greater amount of pigment laterally on the pronotum, but the median emargination of the black pattern is much deeper (figs. 63 and 64). This species, closely allied to the preceding, is recognized because of Dimmock’s statement that the larve differ specifically.

Var. polita Casey.

Having the scarlet color of C. sanguinea (figs. 65 and 66). It will probably prove to be a variety, because in its range, the northwestern States, yellowish indi- viduals are occasionally met with. It is possible, however, that these speci- mens are caused by different degrees of fading.

CEP ere

Fig. 65.—Variation of pronotal pattern in C. munda var. polita male. Fie. 66.—Variation in pronotal pattern of C. munda var. polita female.

3064 b. Cycloneda ater Casey.

Locality unknown. This wholly black species is paralleled only by A. lugubris.

66 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

Genus OLLA Casey.

The color-pattern in this genus is so diverse in its elements and lines of variation from Coccinella that I believe that it is entitled to more than the subgeneric rank which Leng gives it. It has stronger resemblances to Neoharmonia ampla in its color-pattern than to any Coccinella.

3065 a. Olla abdominalis Say.

Distribution: Indiana to Texas, Arizona to California.

The crescentic shape of the subsutural intermediate spot has its parallel in Neoharmonia venusta and ampla and one specimen of H. convergens. The variation in the color-pattern of the pronotum is given in fig. 67.

a b ¢ Fie. 67.—Variation of pronotal pattern of O. abdominatlis. pes ag Var. semilunaris, n. var. The subsutural and intermediate spots of the intermediate row are fused and extended towards the apex to make a bowl-shaped mark (fig. 68). This ex- tension is a unique feature, for it is not an equilateral enlargement nor a confluence of spots, but a determinate extension in one direction. Arizona and Texas. Var. minuta Casey. The elements of the intermediate row and the apical spot are confluent, but with reentrant angles (fig. 68). Brownsville, Texas.

6 c e 2 g : a b ¢c d : e

Fra. 68.— Variation in elytral pattern in O. abdominalis. f, var.minuta; g, h, var. semilunaris, Fra. 69.—Variation of color-pattern in O. plagiata. All from Arizona.

3065 b. Olla plagiata Casey.

This is the melanic derivative of Olla abdominalis (fig. 69) and may be found to be intergenerating. In the meanwhile it is best considered dis- tinct. The Florida specimens differ, Casey says, as well as those from Las Vegas, Mexico, but the difference is slight and, in the lack of a large _series, they will not be described here.

ADALIA. 67

Genus ADALIA Mulsant. 3068. Adalia bipunctata Linnzus.

Distribution: Europe and Eastern and Central States. Typical pattern: Spots 2+3. Double nature of the band not apparent.

Form pruni Weise. Spots 2+3. Double nature of spot apparent. About 1 per cent at Cold Spring Harbor.

Var. herbsti Weise. Spots 2, 3. I have not yet seen this variety in America, but it may be confidently expected because of the occurrence of form pruni and the probability that when found it has been considered to be the A. frigida var. parvula.

Var. rubiginosa Weise. Spots 1, 2+3. Less than 1 per cent at Cold Spring Harbor.

The 17 other varieties given by Weise (1879), nearly all of which show greater amounts of pigment, have not yet been recorded from America with the exception of —

Var. quadrimaculata Scopoli =< humeralis Say? Elytra black, except for a reddish humeral area and a reddish spot near the suture at ?. Widely distributed, but always uncommon.

Fig. 70.—Variation in pronotal pattern in A. bipunctata. c=mode.

Subspecies of North America: The lack of all varieties with the excep- tion of pruni, rubiginosa, and quadrimaculata, is typical of North Amer- ica asa whole. While some of the other varieties will be found, they will be great rarities, and one may say that North America is characterized by the general lack of the dark varieties, excepting the quadrimaculata.-

Burgess (1903) has found the variety quadrimaculata and the typical species to be intergenerating. The heredity was perfectly segregate. The variety quadrimaculata is found in Europe in company with numerous varieties which intergrade toward the typical species and to the almost or wholly melanic variety lugubris. It is singular that with us these other varieties should be absent, leaving only the variety quadrimaculata. It can not be because variety quadrimaculata is a position of much greater organic stability, because in Germany (Schroeder, 1901) it is greatly out- numbered by the allied variety sexpustulata, which is less melanic. It justifies the view that at least some of these varieties are not merely ontogenetic in origin or fluctuations, but true varieties representing inheritable positions of organic stability.

68 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

The pronotum in the dark variety guadrimaculata is black, with only a very light lateral margin which may not be complete. The type and light varieties have a range of variation shown in fig. 70. This exceeds that shown in Europe in the direction of less pigment. The longitudinal median mark fails in only 3 cases in 97. I have not seen any of our specimens with a totally black pronotum, such as is known in Germany.

From 1 impregnated female of variety rubiginosa I raised 8 beetles; 2 were of the typical species and 1 of the variety rubiginosa, with the spot slightly smaller than the parent. The heredity is therefore segregate.

3066. Adalia frigida Schneider.

Distribution: Lapland, Siberia, and United States. Typical pattern: Spots 1, 24-3, 4+-5. Rare. Found by Casey in Colorado and named by him ornatella (fig. 71). Var. immaculata n. var. No spots. Unknown from Europe. Var. melanopleura Leconte. Spoti. Colorado. Unknown from Europe. Var. parvula Weise. Spots 2, 3. Our commonest variety in North America. Var. siberica Weise. Spots 2+ 3. One from Tyngsboro, Massachusetts. Var. fasceta Weise. Spots 1, 2-+3,4-+-5. One from Springfield, Massachusetts. Var. postica n. var. Spots1, 2, 3,445. Two from Springfield, Massachusetts. Var. hyperborea Paykull. Spots 1+2+43,4-+-5.

Var. humeralis Say or n. var. Has the pattern of A. bipunctata var. quadrimaculata. Widely. distributed, nowhere common. It is uncertain whether the A. humeralis Say was an A. bipunctata or A. frigida. Since the name quadri- maculata Scopoli antedates it for the variety of A. bipunctata, I apply humeralis to the corresponding variety of this species.

Weise names 138 other forms showing various other combinations of absence or confluence of some of the 5 spots. Most, if not all of these, will probably be found in the United States eventually. The forms with absence of 4 and 5 constitute a variety as well as those without any spot. The other patterns are forms of the typical species or of these varieties.

The pronotum shows variation (fig. 72) from an M pattern with 2 lateral spots, in which it approaches that of A. bipunctata, to one with pigment so reduced as to reveal a number of elements nearly identical with those of Cleis picta and Olla abdominalis, closely allied genera. The following forms of pronotal pattern may be recognized. None of these are of var- ietal rank and the amount of pigment has only a low correlation with the amount on the elytra.

Form A. The M-pattern so reduced that the component elements are readily recog- nized. The lateral spots separate.

Form B. The M-pattern apparent, and resembling that of A. bipunctata. The lateral spots separate.

Form C. The M-pattern heavier, with lateral spots narrowly united.

Form D. The M-pattern much extended, with lateral spots broadly united, the pigment encroaching so as to leave 2 light spots at caudal margin instead of 1.

70 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

Subspecies of North America: The typical condition in Europe is not common with us. The variety parvula far exceeds it in numbers.

The case of the variety humeralis is especially interesting. In the East- ern States where A. bipunctata is found, many specimens with this pat- tern belong to A. bipunctata, as is known by the cross-breeding obtained by Burgess. The humeralis pattern is found, however, throughout the Rocky Mountains and Pacific States, where A. frigida and annectans occur but not A. bipunctata. In fact, this pattern is more abundant there than in the East. Since the two species are closely allied and each is known to give many variations in the melanic direction which show many analogies, I believe that this pattern exists as a variety of each. It becomes neces- sary to consider these western specimens as variety humeralis of A. frigida or variety humeralis of A. annectans.

TABLE 18.—Heredity in Adalia frigida.

Offspring. Mother, variety. | No, | ; | H I , Gove yper- mmac- ee Parvula. | Sibirica. | Postica. Faceta. | Dorea. lata |--_——- | | | 7 Spots present.... | | 2,8 24+8 ree ae | 1 q 2+3,/ No spots NEE ye ae eae eens ee aes re eet eRe he | 7 | M2 WARES * MRSS 1 2 i er teal te PPP MSP se Po weeese apes ipekestz 1

Three specimens that were kindly sent to me by Mr. G. W. Dimmock, of Springfield, Massachusetts, that proved to be impregnated females with spots 2 and 3 only, gave offspring as in table 18. It will be noticed that where spots 4 or 5 are present at all they are well developed, thus consti- tuting a unit-character.

AE CAD EW ED ie ew

Fr14@. 72.—Variation in pronotal pattern in Adalia frigida.

3063. Adalia annectans Crotch.

Distribution: Western States and British Columbia.

Lack of intergrades and of breeding experiments make it desirable to retain the specific status of this pattern for the present, in spite of its close relation to A. frigida. (See fig. 71.)

Typical pattern: Spots a, 1, 2, 3, 4, 5, 6, 7. The numbering is that of Weise, with 6 the humeral spot, and 7 next it.

Form A. Spots 1 + 2, 3, 4, 5, 6, 7, Form ovipennis Casey. Spots 1, 2, Form B. Spots 1 + 2, 3, 4, 5, 6, 7, Form C. Spots 1 +3, 2, 4, 5, 6, 7, Form D. Spots 1 + 2, 3, 4, 5, 6+ 7, 4. Form E. a, 1-+ 2, 3, 4+ 5, 6+-7, 4. Form F. a, 1+ 3, 2,4+ 5, 6+ 7, 3. Form G. 1, 2, 3, 4, 5,6+7 +3. Form H. 1+2-+5, 3, 4,6+7-+4. Var. transversalis Casey. Spots 1+2-+3,4+5,6-+ 7, a. Form I. Spots (1 +2+3) + (4+5),647+a. Form J. Spots 1+2+4+3+5-+ suture, 4, 6+ 7, a. Form K. Spots (1+ 2+3)+ (4+5) + suture + margin, 6+ 7. Var sexpustulata n. var.~A. bipunctata var. serpustulata. The black humeral spot within the reddish humeral area places it as variety of this species rather than as A. frigida. Hood River, Oregon. Var. ocellata n. var. Specimens with the black pigment revealing 2 light spots on the dark ground instead of 1 as above. Hood River, Oregon. Var. humeralis n. var.=<>=A. bipunctata var. quadrimaculata. The black humeral spot within the humeral reddish area place it as a variety of this species rather than A. frigida. Widely distributed.

Most of the species of the dark varieties given above are in the Schwarz

| collection of the U. S. National Museum, and were collected at Hood River,

Oregon. They seem rare elsewhere, but this may be because they were more diligently sought there.

The pronotum of this species shows variation closely parallel to that of A. frigida. So far I have seen only the patterns given in fig. 72. The other patterns of A. frigida may, I believe, be looked for in this species when larger series become available.

The legs of this species (at least of its light varieties) in Spokane County, Washington, are uniformly yellowish-brown and there is an area of the same color on the ventral surface of each segment of the abdomen, which in A. frigida is nearly black. Larve brought from Washington to Cold Spring Harbor emerged here with the western color. None were raised in the East, however, from the egg.

72 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

Genus CLEIS Mulsant. 3069. Cleis picta Rand.

Distribution: Northern United States and Canada.

This species is certainly not congeneric with many of those placed in Harmonia by Mulsant, as shown by the pronotal pattern, which shows much more affinity to Adalia. The pronotal and elytral variation as shown in figs. 73 and 74 is great and apparently continuous.

J SAUDE

Fia. 78.—Variation of pronotal pattern of Harmonia picta. b=mode. Fia. 74.—Variation in elytral pattern of Harmonia picta. e=mode; b, c, from Tyngsboro, Massachusetts; d, e, f, from Wisconsin; g, h, from New Jersey; i, from New Mexico.

Var. blanchardi, n. var.

Elytra without black pigment. Throughout the range. Named in honor of Mr. F. Blanchard, who has kindly called my attention to the fact that this is not a sex difference. The former view that this spotless pattern was that of the male arose from the fact that it is much more numerous among the males.

Form A.

Pattern reduced to a few small marks, as in fig. 74, a and b. Two specimens from Tyngsboro, Massachusetts. It is improbable that this form is limited to Massachusetts. Similar vigilance should reveal it elsewhere.

Form minor Casey. The costal spot at ? confluent with the vitta and margin (fig 75). Form contexta Mulsant. Basal band continues from vitta to suture. Form concinnata Mulsant. The elytra nearly black.

P22) DPE

F1@. 75.—Elytral pattern of 3 Western specimens of Harmonia picta. a, from Paradise Park; b, from Washington; ¢ from Vela Pass, Washington. .

Fra. 76.—Variation in pronotal pattern in Harmonica hudsonica. a, from Carrs Peak; b, from Falmetz, Arizona; c, from Whitefish Point, Lake Superior; d, from Marquette, Michigan.

Subspecies of Pacific Coast characterized by form minor and intergrades to it. Subspecies of New Jersey characterized by the presence of the very

dark patterns. 3069 a. Cleis hudsonica Casey.

Distribution: Boreal Zone in Canada, Wisconsin, Michigan, Mount Wash- ington, New Hampshire, and Black Mountain, North Carolina.

The variation, while considerable, as shown in figs. 76 and 77, is appar- ently continuous. It does not intergrade with C. picta, for the pattern is made up of different elements.

AGRABIA—ANISOCALVIA. 73

Genus AGRABIA Casey. 3070. Agrabia cyanoptera Mulsant.

Distribution: Mexico, New Mexico and Arizona.

While the peculiar blue color is something quite different from its allies, the pronotal pattern (fig. 78) points to a relationship with Neoharmonia. It is probable that its blue color is a feature of relatively recent acqui- sition.

DPI eomp

Fic. 77.—Variation of the elytral pattern in Cleis hudsonica. c=mode; a, d, from Palmerly, Arizona; 6, from New Hampshire. Fig. 78,—Variation in color-pattern in Agrabia. Atl from Huachuca Mountains, Arizona.

Genus ANISOCALVIA Crotch. 3072. Anisocalvia duodecimmaculata Gebl. (Fig. 79.)

Distribution: Lake Superior, Hudson Bay, British Columbia, and Siberia. Form elliptica Casey. The two spots, transversely placed at middle of the elytra, confluent.

73 b e d c f g h a 7 k l m n

Fie. 79.—Variation of elytral pattern. Anisocalvia duodecimmaculata: a=mode; a, from Oregon: by, from Hudson Bay. Anisocalvia quatuordecimguttata: h=mode; ¢ , d, from Montana; e, from Mar- quette, Michigan; /, from Catskill Mountains; g, from Canada; h,j,k,n,from British Columbia; i, from Hudson Bay; /, from Massachusetts; m, from Hood River, Oregon.

3071. Anisocalvia quatuordecimguttata Linnzus.

Distribution: Northern United States and Canada.

There is in the Museum of Comparative Zoology a very large series of the beetles of this species, or at least of this genus, collected by Mr. H E. Morse, at Lake Champlain. At the time of my visit to the museum they could not be found. Until this much-needed series can be carefully studied a revision of this genus is inadvisable. All the patterns are there- fore given in fig. 79, with no attempt to define their status. Some are said to exist in one sex only, but this requires investigation.

74 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

Genus ANATIS Mulsant. 3075. Anatis quindecimpunctata Oliv.

Distribution: Northeastern United States and Canada.

The dark form mentioned by Leng may well be an age difference, so much does appearance change in specimens when kept alive. There is, however, variation of spots (figs. 80 and 81).

Form A. Inner apical spot absent. Form B. Scutellar and subsutural intermediate spots confluent.

> @ © a b c Knap Lo) a b c d €.

Fra. 80.— Variation in pronotal pattern in Anatis quindecimpunctata. b=mode. All from Stony Lake, Michigan,

Krae, $1.—Variation of elytral pattern of Anatis mali and quindecimpunctata. aand b=o mali. All from Stony Lake, Michigan,

3075. Anatis mali Say.

Distribution: Michigan, Indiana, Wisconsin, Idaho, and Washington.

The variation of the pattern (figs. 81 and 82) seems to be continuous. There are intergrades of A. quindecimpunctata towards the two costal spots of A. mali. The pronotum and shading of elytra, however, distin- cuishes it sharply from A. quindeeimpunctata. Its range is within that of quindecimpuctata, which is probably the parent species.

4 b c d a b €

F 14. 82.--Variation of pronotal pattern in Anatis mali. a=mode, All from Stony Lake, Michigan. F1iG. 83.--Variation of color-pattern of Anatis leconiter. From Southern Arizona,

3076. Anatis rathvoni Leconte.

Distribution: California and Oregon. Form A. Withsome or allof the spots placed as in A. quindecimpunctata indistinctly indicated. 3076 a. Anatis lecontei Casey.

Distribution: Colorado, New Mexico, and Arizona. The black limb is variable in extent (fig. 83). Form A. Without the basal spots on the pronotum. Throughout the range.

EPILACHNA. 75

TRIBE EPILACHNINI. Genus EPILACHNA Chevrolat. Epilachna borealis Fabricius.

Distribution: Eastern United States. Type: Spots 1, 2, 3+ suture, 4, 5, 6-+ suture, 7 (fig. 84, 5). Form spots 1, 2, 3-+suture, 4, 5+ 6-+ suture, 7 (fig. 84, d); 6 in 156=4 per cent of females and 27 in 214 = 13 per cent of males at Cold Spring Harbor, New York. Form spots 1, 2, 3 + suture, 4, 5, 6, 7; 3 per cent at Cold Spring Harbor (fig. 84, a). Form spots 1, 2-+.3 + suture, 4, 5, 6+ suture, 7; 2 in 370 = 1— per cent at Cold Spring Harbor. Form spots 1, 2+ 3-+ suture, 4, 5 + 6 + suture, 7; 1— per cent at Cold Spring Harbor. Form spots 1+ 2+3-+ suture, 4, 5, 6+ suture, 7; 1— per cent at Cold Spring Harbor

(fig. 84, e).

Form spots 1, 2, 3+ suture, 4, 7+5+6-+ suture; 1— per cent at ColdSpring Harbor fig. 84, g).

Form spots 1, 2, 3+ suture, 4, 6+ suture, 6+ 7; 1—per cent at Cold Spring Harbor (fig. 84, f).

eG

Fra. 84,—Variation of the elytral pattern of FE. borealis. b—mode; a tog, from Cold Spring Harbor, N. Y.

See he

Fre. 8.—Variation in the pronotal pattern of Z. borealis. e=mode. ato /, from Cold Spring Harbor, New York.

Although not a color variation, 1 specimen with an emarginate pronotum cephalad seems worthy of mention here. It did not appear in any way pathological or traumatic. Intergrades are so numerous and the inherit- ance of these forms is so imperfect that they do not deserve varietal] rank. No geographical differences are known in the range, except that it has given way to its derivative, H. toweri, in Texas and vicinity. The pro- notum shows a wide range of variation in the amount of pigment (fig. 85), which the modification and heredity experiments given later show to be largely the result of modification. There is, however, one distinct variety.

Var. A. Lacks lateral pronotal spots; 1— per cent at Cold Spring Harbor (fig. 85, a).

76 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

MEASUREMENTS.

The polygon (fig. 86) shows that the confluence of spots 5 and 6 does not represent a well-marked center of variation. Fig. 87, which is intended to test the same matter for the confluence of spots 2 and 3, fails, because of the rarity of this variant, to show its true status. The decided skew- ness in both polygons will be discussed in another connection.

(a) The distance between and confluence of spots 5 and 6.

(b) Males.

{ (c) Females.

lA (a) All individuals.

c a i A : pe Spe (a) 23 22 212019 18 1716 151413121110 9 8 7 6 5 4 8 101234 5 6 78 9 10111213 14151617 1819 2021 22 23 (010001 1 4 5 15121518272413 8 11665 §32560211112021000010101 O—214 (e) 00323 4 64 11121418M1210103 4055 w,2929200001000001000000 0 0-156 (@) 0132235 7 8 17°27 27°28 32 39 34 2311151211 1014°8"4 276221112202 101001010 1 03874 Fra. 86.—Variation of the distance between and confluence of — 5 and 6 in Zpilachna borealis at Cold

Spring Harbor, New York. MODIFICATION.

I was unable to modify the elytra by subjecting the prepupa and pupa to changes of humidity and temperature. The pronotal pattern, how- ever, is highly subject to modification by high temperatures. Fig. 85 gives the range of the wild specimens. Individuals kept in an incubator at 40° C., and even those placed in the hot attic of the station buildings during summer months, became very light in every case, as shownin fig. 88, and closely approached the Texan species E. toweri. Some of the spots, besides being smaller, have a lesser amount of pigment proportionately, so that they were in some cases brownish rather than black. Prolonged refrigeration, darkness, increase and decrease of humidity produce no appreciable effects. Experiments with the high temperature referred to, combined with high humidity, gave results similar to those with the high temperature alone. Excessive heat was, therefore, the only agent to produce modification and this was in the direction of reduced pronotal pigmentation.

The pupal coloration is subject to high pigmentation after a temperature of 17° C. With the temperature at 40° C. the exuviz possess no pig- ment, except in the spines. In the former case, the pigmentation was generally associated with lines of greater frequency of spines.

EPILACHNA. 77

HEREDITY.

Since it is difficult to get two broods through ina year and since eis loss in hibernation is heavy, the experiments were frequently interrupted at the end of one generation (table 19). The death-rate from hibernation of the forms with confluence of spots was higher. This matter should be

(a) The distance between and confluence of spots 2 and 3.

(b) Males.

(c) Females.

(d) All individuals.

3 4 2 6

Fig. 87.—Variation of distance between and confiuenee of spots 2 an borealis at Cold Spring Harbor, New York.

experimentally tested with large numbers. The heredity of the forms with confluence differs from that of the varieties of Hippodamia with con-

fluence, as would be expected where intergrades are morefrequent. The forms are inherited as if they represented an extreme. in fluctuation, but

78 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

with this difference, that the confluence will crop up in some of the progeny nearly if not quite to the full degree. Even though there is seg- regation of the germinal representatives of these characters, fluctuation is so great as to obscure it. The explanation of these facts on the basis of blending heredity seems at first more satisfactory. The difficulty with that hypothesis is the failure of these variations to have long since disap- peared by swamping. We may assume that variations in the direction of confluence are determined and that they continue to appear from time to time. Their failure to carry the whole species to a greater degree of con- fluence is probably checked by selection either in hibernation or through predaceous foes or by subponderance.

TABLE 19.—Heredity in Epilachna borealis. {C,S. H.—Cold Spring Harbor. ]

| Father. | Mother. 2 and 3. | 5 and 6 5 and 7. | 6. | » P ee FETA ‘ | i] | A= . | Le | =>} | = Oft- iP Of |\No.) Nor. 3 2a |S, | ey 3 a z By No. | spri No. spring | S12 \~aes |e] | 7 Oo. spring | o spring zis +/ 338 Sie it+ieflei+ 3 “35 | | |BIS| | BBP] 2/2] 1818) |g8|Be S | | <4|2|} |n8 2 |D a |D re | —— a | aoe ee | | | | | / | | C.S, H.| 0.8. H.| 243 ¢ 10.8. H.! 38 | 80 |...) ..1.. | Pe 7h Beco ee Jy Bet pe | 479.3 | 243 | a7) y| 248 ire ela net eee lhe Bl eee se | C.8. H.| C.8. H.| 237 ¢ | 0.8. H.| 17 | M184) .. |. |... | sss |e | BE] Sah pe elke ; | C.8. BOSH mee (OS BY MY Pee cad. Boke ce tice Fotee cl alge ih |C.S, H.| C.8. H./ 236 ¢ C.8.H.) 24 | 193 | 2. |. Yee re ee | ee i 4 | 2762 | 236 |276¢| 208 | 50| Mad! ..|..|..] 1 | @ 18H .. | 2] 6c). a | 848 3 | 248: |8489/ 243 | 5] 2 |...) *s1*s| FM | 25]... | .. ar oes | | 2012/¢C.8.8./91°/0.8. HB.) 1) MI] ..] 0] .. | A ie Reece iv fee . | 3002 |C.8.H.|300¢) 287 | 11 | M54 | 5 oe, 2 5 es a, | |C.8. H.|C.8. H.| 4709'C.8. A.) #8] 22°)... ).. 1.) .. [Bal tal dé | 8) ..1M . | C.S, H.|C.8.H-|216¢|/C.8. He) 2) foi We. |My tn.) pay cubes | ee a 'C.S. H.|C. 8. H.| 267¢/C.S. H.| 24] 103 | i i | $| 34 |Ma4l ta|/M|.. a (CS. H.| CUS. H| 4709 O.8. He) 28 | 2 eed Boe ee %/1 |..| 3)4 |M | | | } | i

|

* The individual with 2+ 3 was 1+ 2+ 3 on one side. t Pigment along suture towards base in one.

t Both having 5 ap. 7 had 5 ap. or + 6.

2 Pigment along suture towards base in one.

Epilachna toweri n. sp.

Distribution: Texas (Comal County).

This species most nearly resembles E. borealis Fabricius. The pigment area is reduced, but along the lines of variation seen in #. borealis. The space between the two median spots on the pronotum is about one-half the length of the pronotum. The caudo-mesal spot is free from the caudal margin of the pronotum. The lateral pronotal spots are about one-fourth the length of the lateral part of the pronotum. The mesal spot of the intermediate row on the elytra is not confluent with the suture. So faras color-pattern goes, these beetles would not be given more than subspecific status. There is, however, a difference in form that is distinctive. The elytra are more convex, increasing the ventro-dorsal dimension. In the basal part of the elytra the slope is only slightly curved, as shown in fig. 89. The area between the two humeral spots is thus flattened and the

EPILACHNA. 79

height of the elytra in the basal region is low. The elytra are also a little broader and the pronotum a little flatter. The ground-color is a lighter yellow than that of E. borealis. It is desirable to find if this species does not feed upon some other food plants than those of E. borealis. This species is named in honor of

Prof. W. L. Tower, of the University ) ge xg of Chicago, whose experimental work :

. ° ° a c upon the genus is awaited with great interest.

‘Epilachna corrupta Mulsant. Distribution: Arizona, Mexico, and New Mexico.

Typical pattern: Spots 1 and 4 are z free from the suture and in place of a large single apical spotare 2 smaller *"“ishacnas boreutis exposed tote ones placed well apart pee sae oi teeinad oberon comet fee pike eatin th

Form A. E. toweri; dotted line, E borealis; b, color- pattern of pronotum and elytra. Mesal and intermediate pedis of the inter- mediate row narrowly confluent; 1 specimen and some intergrades. _ Var. juncta, n. var. The two apical spots confluent; 1 specimen.

The food plant is the bean. The larve refused to eat cucurbit leaves of several species. They were successfully reared, however, on bean leaves at Cold Spring Harbor. The imagoes showed no modification there in the direction of the pattern of E. borealis; in fact, a few showed a slight decrease in the amount of pigment.. The attempt to hybridize them with E. borealis was unsuccessful.

Epilachna mexicana Guerin. Distribution: Mexico and adjoining parts of the United States. This species shows a reversed pattern, light spots upon a dark ground. Living specimens could not be obtained nor enough dead ones to make a study of the variation.

80 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

ce

Fie. 90.—Varieties of Hippodamia convergens, arranged to show the lack of straight lines of divergence.

ie i ia

PART III.

GENERAL DISCUSSION. VARIATION.

The variation in the color-pattern of these beetles is so great that the first impression is that it would be useless for taxonomic purposes. Some entomologists have on this ground ventured to say arbitrarily that color differences alone shall be an inadequate specific distinction. While this practice has the warrant that where only a color difference is found, the distinction is generally in fact varietal, such a discrimination against color is of course untenable when the two forms are a distinct unit which do not interbreed. On the other hand, Major Casey considers the color dif- ferences in this family of such taxonomic value that he has created many species thereon which others believe must be ranked as forms or varieties.

After an acquaintance with these beetles in large numbers, the varia- tions are seen to have very different values. A considerable addition or loss of pigment at some places on the elytra or pronotum will be recog- nized as a mere fluctuation common throughout the range, whereas a very small amount of pigment at another point would show it to be a separate

_ variety or even a distinct species. To the experienced coleopterist this is

a commonplace, but the neglect of it by the Wallace-Weismann school of selectionists makes it worthy of note.

In general, the mere amount of pigment has the least significance. If this pigment is not evenly added to the spots, it becomes more significant. But if it is along the set lines for that species, it is of comparatively slight consequence. It is when it breaks out from the set lines to appear at a different point or to extend in new directions that the systematist may lay stress upon it. :

That the variation is determinate is shown in many ways. By determi- nate variation is here meant a progressive variation in some definite direc- tion, originating within the germ-plasm either with or without external influence, and gradually, by waves or by leaps. The progeny differ in the long run from the parents in some one direction, instead of offering indif- ferently plus(+) and minus (—) variations in all the old characteristics and having new characteristics in all directions, as seems to be assumed by Wal- lace and in the earlier writings of Weismann. I have not used the term orthogenesis, as I think that it is generally used for determinate evolution operating unswervingly and without divergence in one direction for a long period. Eimer thus used it, contending that the longitudinal pattern steadily and unfailingly develops intoa transverse pattern. It is not impossible that such phenomena exist, but the determinate variations here described can not be so characterized. Diversity prevails. Determinate variation will

be working in several features at once, advancing now here and now 81

82 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

there, in one direction or the reverse direction. In fig. 90 the variations of Hippodamia convergens are arranged to show this. In fig. 11 they were arranged merely to show the successive steps of the several lines of devel- opment. There is no kinetic or progressive principle here at work causing a general march. Rather, these beetles have a repertoire of development- lines which are called into activity here and there as the environment sets one or another into play. This repertoire variation is illustrated in table 20.

The one variation that has appeared most frequently is spotlessness. I believe that this is because loss as such is an exigency to which determin- ers in the germ-plasm are more subject in the long run than any other; just as all kinds of machinery have this one frequent cause of disablement in common, the loss of some part. This item in the repertoire of varia- tions is, then, common to most characteristics and organisms. When an organ is composed of many cooperating parts, such as an eye, this liability to loss or deficiency of one or another feature of its make-up renders it more liable to variations of defect than a characteristic which is simple in

TABLE 20.—Repertoire variation in the color-pattern of the Coccinellidz. {For key to table see page 83.]

| | | eg | ag Pie A 3 | Beramuser tia seule ‘ aie | ro |ldo! gs Po) ‘ ; Genus. 3 % z | | Bs z we | 38 3 eal =| = 22 ae 12 |e) 2 (se /ee! Ss wel] e | e212 Ss S > | g | 5 35 | el. & 2 | 5 2 | a) 8 | 3 | S/S (es1es]/ o 1s | sj) -3 n | D ;me |e | mo le _ Z |e Oo 1m o | | | as er pars Saag | \— asd ee a Anisostictu........ la 2h bee 3 | Naemia........... 2a 2b |} 2 : Macronaemia..... cand! | comghf:<asep dee eae | MH Paranaemia, ..... re reef Bests hee asia te ie 4 Megilla............ Ae Fick. wasn) RUGS ce aula ages DB 14e ..adkd pha ae sae Sem ON Clade od LiaeaeH Ceratomegilla,.... ene Mere eres emery Is Woah ye iocen) oye ky ye et ay Pa eae ais'oy 6) Worigueia! 20065 oo ere Poe 7 STE MeN ree ey lee sy) a fe Gee eS ERP a ae Hippodamia...... 8a ee ee en eee 12 8 10a |) lle | iz... 9e 10¢ | .--. | 9 la 9c 10b 20 } 10 | 9a 15 19 ll | 13 / 19 14 | 20 15 18a 16 21 1lb 17 18 Neoharmonia..... 22 23 HA 24 = Coccinella,........ re 5S rae Es Wee eee PL) ee ag Ry a 2 a Jia 29 28 | 27 30 31 82 33 Cycloneda......... - SUL ant brake JL) ON ARR Ane re SG ae Me oe Te ete) ec: 7) Ra POBMBA 6.5 ies Sees preg Pe OO ae eee 401 40° |. :,..s'|\ sane. 400 dla 4lc} 4l 42 42a | 42b Harmonia......... Jone |) ABR dec | cece th] Sane) | eee) 4B) | AB MBIT. ey ei | ad Lan Agrabia..........: EM ey Pe ree ee ieee eres) SRM rec 4 RN (ee 45 Anisocalvia ...... 46 oat Ol) | ETRE Sr: cate Pp seeks w Wares 2.) ib oe AP Se Ase i aren AONB SE 36 <5 cake 48 EPs on A gees: | cake, | ambien ds gal) eke Anes sb ewe as 51 | 49 Epilachna......... 52 5d 52a 53 54a 54

GENERAL DISCUSSION. 83

its structure. So, then, determinate variation in the direction of loss is widespread. That its evolutionary effect is not greater is the result of natural selection, because so great a difference as actual absence is more likely to have selection value than small differences, and because it is often opposed by the activity of determinate selection towards some constructive end. Thus in Hippodamia convergens variations of loss probably arise in regions where determinate variation is building up a pattern of confluence. Such negative variations would serve merely as a slight and easily sur- mounted obstacle.

This hypothesis makes it unnecessary to resort to such questionable hypotheses as use-inheritance or panmixia, to explain the decadence of unused organs such as the eyes of cave animals. If use-inheritance or panmixia be accepted to explain these cases, then no useless organ should

persist. This persistence of some useless organs, such as the anal spine

in some moth larve, offers no difficulty to the hypothesis of loss of deter- miners, for the frequency of this loss would depend on the nature of the determiner-complex and on the presence or absence of opposed lines of determinate variations.

It is not uncommon to find two spots which are confluent in some speci- mens showing in others, where there is no confluence, pointed extensfons each directed toward the other. Such may be seen in many of the figures. It is abortive confluence. This indicates that confluence is not the pig- mentation of a third unit, a connecting bar, between the two spots in question. It is rather an extension of pigmentation from the spot in a

9. convergens. a. Var. obsoleta. b. Var. moesta. e. Var. quinquesignata, d. caseyi, ete. e. Var. A. 10. spuria. a, Var. com plexa. b. — ania. ts 5, 6.

ec. Var. californica.

Anisosticta. Hippodamia. Adatlia. 1. E lo. cockerellii 40. i en ange é _ 4@, Var. bitriangularis. 16. dispar. - Var. herbsti. Naemia. i7. parenthesis. b. Var. quadrimacu- 2. seriata. 18. apicalis. lata. a. Form A. a. Var. Ti ee c. Var. b. Var. litiginosa. +5+6+4. 41. frigida. Macronaemia. 19. americana. a. Var. eg tome 3. episcopalis. 20. sinuata. 5. Var. immaculata. Paranaemia. 21. falcigera. e. Var, humeralis. 4. vittigera. Neoharmonia. 42. Annectans. Megilla. 22. ampla. a. Var. humeralis. 5. maculata. 23. venusta. b. Var. transversalis. a. subsp. floridana. 24. notulata. Harmonia. Ceratomegilla. Coccinella. 43. picta. 6. ulKei. 2%. johnsoni. a. Var. blanchardi. Eriopsis. 26. perplexa. b. Var. concunator. _ 7. connexa. a. Var. subversa. 44. hudsonica, a et 27. novemnotata. Agrabia. glacialis. a. Var. franciscana. 45. cyanoptera. a. Var. G. 5. Var. confiuenta. Anisocalvia.

4§. duodecimmaculata. 47. quatuordecimguttata. a. Var. 6. Var. Anatis. 8. eemarnnia. 49. mali 50. rathvoni. 51. lecontei. ilachna.

52. borealis. a. Wes. 28+ ts +6+»,

Ep

84 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

definite direction. These extensions in the case of most spots are always true in their aim and meet each other exactly. In others, as m and a in Coccinella, the approach is ragged. In those beetles in which the black spots are all confluent one would expect, if confluence were the mere result of overgrowth of pigmented areas, that the light areas thus inclosed and remaining as clear spaces in their turn would have projecting points and concave sides. Such spaces are not wholly unknown, they are illus- trated in one variety of Neoharmonia notulata. But generally such lacunz are rounded, giving rounded spaces with a black background. There must, therefore, be in this case a confluence by definite mutual extension of the spots. .

These orderly extensions point conclusively to determinate variation. But since there is nothing in the structure of the elytra to explain why a spot should be here or there in the cephalo-caudal direction, it seems prob- able that the nature of the determiner-complex for the elytra must be such as to cause variation in these particular set lines rather than haphazard.

Asymmetrical specimens are not uncommon in looking over large series and are frequently sufficiently marked as to leave no doubt that they are the result of some inequitable distribution of the determining agent, what- ever it be, in development.

That there are both fluctuations and mutations present in these beetles is probable. But under the name of fluctuations we group three different things. First, there are the modificational differences. Second, varia- tions arise in spite of similar germ-plasm and similar environment, from the difficulty of constructing a soma according to the germinal specifica- tions even with a similar environment. The developmental processes are not adequate to the task of producing exact fac-similes. Third, there are the variations resulting from slight germinal differences. Of these three kinds, only the third is inheritable. All three are to be found in these beetles. Their evolutionary significance is obviously very different.

MODIFICATION.

When a full-grown larva of certain species is subjected to certain influences which are maintained through a part of the pupal period, mod- ification of the color-pattern results. But the portion of the color-pattern which responds differs not only with the influence employed, but also with the different species. Thus 40° produces no noticeable effect upon Cycloneda sanguinea, Coccinella novemnotata, Megilla maculata, and Hippodamia convergens, nor upon the elytra of Epilachna-borealis. But the pronotum of the latter has its pigment very much reduced; in fact, in some speci- mens, to the degree of reduction found in the Texas species E’pilachna toweri. Refrigeration does not affect the pigmentation of the elytra, but increases the pigment of the pronotum, though not to a degree comparable to the reduction of pigment by heat.

GENERAL DISCUSSION. 85

In the Hippodamia convergens, on the other hand, both pronotum and elytra have the pigment increased by the same degree of cold and unin- fluenced by the same degree of heat. The use of various other influences on these and other species produces no effect. Modifiability of this or that feature is then itself a characteristic. It is a characteristic some- times gained and sometimes lost, sometimes always firmly coupled with the feature in question, and sometimes never possessed by it.

The result has been spoken of as increase and decrease of pigment. We have seen that this may not apply at the same time to both pronotum and elytra. Even in one organ, while modification manifests slight influences merely in the diameter of the several spots or marks, with a stronger influence the additional pigment is laid down in certain directions rather than in others. The result is that the pattern is not merely darkened; it is changed.

The interesting question arises, Are these changes identical with hered- itary changes in phylogeny or something quite different? Some are clearly identical, such as the lateral process of the pronotum of Hippodamia con- vergens, and some are clearly different from any hereditary condition known, as shown in fig. 26.

If we examine these two categories carefully, a generalization is possi- ble. The influence of structure is frequently obvious in characteristics which are the result of modification, much less frequently so in hereditary conditions. In modification, pigment follows the veins and the margin to a much greater degree than in inheritance, where some characteristics are wholly free from the influence of somatic structure, such as the spot g. Of course this is just what should be expected. It is the opposite condi- tion, where the hereditary and modificational changes are the same, sev- eral cases of which have been shown in Part II, that calls for discussion.

The most acceptable explanation is that the directing protoplasmic elements in the developing somatic cells have so much in common withthe protoplasmic elements in the germ-plasm that both are similarly affected. If we assume this, some important corollaries follow. An environmental influence may affect the soma and the germ-plasms simultaneously, so that in the following generation there is the same effect as would have taken place if there had been an inheritance of an acquired (somatogenic) char- acter. Such phenomena as this are, I believe, not uncommon, not only in the evolution of these beetles, but very generally. Indeed, recent nco- Darwinian writings are making gfeater and greater concessions to this view. The advocate of the inheritance of acquired (somatogenic) char- acters must show the inheritance of a character which is incapable of affecting the germ-plasm in the same way asitaffectsthe soma. Theseare particularly the kinetogenetic characters. The neo-Lamarckian may well contend, however, that the admission that the germ-plasm is influenced by the environment in such a way as to produce definite evolutionary

86 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

results constitutes a far-reaching compromise which goes far to justify the faith he has placed in the réle of environment.

The fact that some characteristics are never inherited, but are always mere modifications, and that others are wholly free from the influence of modification, has a practical bearing to the systematist and to the breeder. The systematist would do well to determine experimentally, where it is feasible, which characteristics of those he deals with are especially sub- ject to modification, by the differing influences of season and range. Such will be his ‘‘poor’’ characters for species distinction and important ones for subspecific distinctions. The breeder will find that characteristics espe- cially subject to modification must be subjected to as uniform conditions as possible, whereas this effort may be saved with other characteristics.

The modifiability of Hippodamia convergens suggests that some of the sub-species in this family might be merely ontogenetic in origin, but this is doubtful, for all that I have brought to Cold Spring Harbor and bred there have retained their characteristics. The variety moesta would naturally be most subject to suspicion on this score, but it is found on the outskirts of its own range in company with beetles which are even spotless and yet without any loss of its typical character. Any rigorous attempt to modify the germ-plasm, as accomplished by Tower (1906) in Leptinotarsa, requires the subjection of large numbers under elaborately controlled conditions. It is not surprising, therefore, that my attempts, which fell short of these conditions, gave negative results when they did not result in death. Still, I believe from the facts of variation and distribution that some such ger- minal modifications as those in Leptinotarsa might be produced, but prob- ably in a much smaller percentage of cases. Our interest here is primarily in the imaginal coloration. Yet the far greater modifiability of the pupal coloration, when the prepupa is exposed, is worthy of remark. The pupal exuvie of Epialchna borealis varied from colorlessness, except in the spines in the heat experiments, to a highly pigmented pattern of stripes and other marks in the cold experiments described.

DISTRIBUTION.

If the view here entertained of the importance of the direct action of the environment is correct, the task of unraveling the dispersal of a species becomes more complicated. The criteria of Adams and others would be less reliable in their application. The four criteria acceptable to Tower (1906) of the considerable number that have been proposed by which the center of dispersal can be located are as follows:

(1) Location of greatest differentiation of a type (Adams).

(2) Continuity and convergence of lines of dispersal (Adams).

(3) Location of synthetic or closely related forms (Allen).

(4) In some cases, location of dominance or great abundance of indi- viduals (Adams).

GENERAL DISCUSSION. 87

,

By the first criterion the Pacific States would be the center of origin of Coccinella trifasciata, but this seems to me the least probable center of origin. In all other localities this wide-ranging beetle is constant to a pat- tern which is present in other species. In the Pacific region it yields to the general tendency there and gives a series of variations towards reduction. The conditions of the criterion are thus fulfilled, but the conclusion does not follow. The position of greatest differentiation is not only a result of time, but also of the diversity and efficiency of the environmental influence.

The fourth criterion of great abundance is so obviously in the control of the environment and subject to such rapid changes that it is not of aid here.

Multiple origins must be reckoned with. When the environment is capable of producing changes of the germ-plasm, we may expect similar changes to take place independently throughout the one unit of environ- ment and in other disconnected but similar environment units. Hippoda- mia convergens would probably give rise to var. quinquesignata wherever it encountered mountains in the same general region.

There is a great wealth of varieties of Hippodamia convergens in the Western States in contrast to a uniform monotony in the Eastern States. The criteria of dispersal would lead us to conclude that the Western States were the center of dispersal, but I think this is not justified. H. conver- gens “‘flies to pieces’’ in that region. It would do so at the time of origin if it arose there or at time of entrance into it, if it arose elsewhere.

We are, then, reduced to a very cautious use of criterion 3 with slight aid from criterion 2. The results are not sufficient to make it possible to reconstruct the dispersal of any species from those here studied, and only one reasonably safe conclusion appeals to me.

Coccinella novemnotata var. menetriesa and C. transversoguttata are com- mon to eastern Asia and western America. There is a close relation be- tween Hippodamia amoena of Siberia with Hippodamia apicalis of Western America and between Coccinella mannerheimii and Coccinella tricuspis and between Cocinella interrogans and Cocinella sinuata of Asia and America, respectively. Species common to Europe and North America are found in Siberia as well, with the possible exception of Adalia bipunctata, which Leng suggests was introduced into North America. These considerations would lead us to believe that the latest route of dispersal between the Palearctic and Nearctic realms was by the Aleutian or Behring route. I do not believe that we are justified in concluding at present in which direc- tion the greater amount of dispersal took place.

One expects the different species in a region to occupy different niches in the environment. This at least is a corollary of the current belief that every species is as common as it can be, its numbers being limited only by its food-supply, a belief which is the result of the strong Malthusian lean- ings of Darwin. The major species of the coccinellids do not seem to be

88 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

so distributed. With certain exceptions which we have given, the species of Hippodamia and Coccinella are in quite general competition. They are characterized for the most part by very wide distribution and extensive overlapping of other species. With the exception of a few species like Cleis hudsonica, there is little zonal distribution. The apparent zones of distribution, as given for New Mexico by Cockerell (1898), are in marked disagreement with other regions.

The abundance of the several species is also very erratic from place to place. At Cold Spring Harbor the order of abundance of the several species differs from year to year and is quite different from that given for other localities not far distant. The conclusion seems evident that the exigencies of the death-rate are great and that these species have diffi- culty in leaving 2 progeny for each pair, not because of overcrowding, but from some other unfavorable circumstance. Small considerations, such as the activity of some disease of aphids here or there, turn the balance against this or that species, though it may be in the middle of its range. If this be the case, then just such wide overlapping ranges should be found. The distribution of the varieties as opposed to that of the species must be separately considered, for here we have different conditions. The facts of the varietal distributions apply also to some species of a narrow range which are closely related to some species of wide range and which are more or less certainly derived from them at a relatively recent date. These are young species not far removed from varietyhood and, not unnat- urally, distributed much as varieties are. They will be distinguished as minor species in distinction from major species.

In Hippodamia, especially, we have wide-ranging major species. Some- times on the outskirts, sometimes on the interior part of its range, are found the lesser ranges of the varieties and minor species. This is illus- trated diagramatically in fig. 91 for Hippodamia convergens and its deriv- atives. Such a distribution is, I believe, very significant in the study of evolution in this group. It points strongly to the directive influence of the environment and to the réle of segregate heredity in the evolution of a species up to the varietal stage, at least, within the territory occupied by the mother species. The difficult step by which the variety acquires intersterility and thus becomes a species must be discussed in a later section.

When the new species has arisen, the extension of its range, as we have seen, is not checked by any coccinnellid saturation of the environment, and depends largely upon the degree of success it has in the distribution of its eggs and in its success in hibernation. If relatively successful, it takes its rank as a major species; if less so, it continues as a species of narrow range; if still less so, it continues as a rarity which persists, in spite of its high death-rate, through the mass of recruits constantly fur- nished by the mother species.

°

RE NS NE NS NEN

~,

+ N+ N+

GENERAL DISCUSSION.

F R = : S was SNS OAC

| SON

Peas NISC SONG:

RR, RE SES SS

PES IR Oy ON OANA UN

BESS

—

H. glacialis + |H.bowditchi H. Convergens v. 5-signata H. Convergens v. caseyi © |H. convergens v.moesta H. converg. v.lS-maculata ¢ |H. convergens v.obsoleta

Fig. 91.—Distribution of some derivatives of Hippodamia convergens in the United States (diagrammatic),

90 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

David Starr Jordan has proposed a law of distribution of great value, not because it is universal, for it is not, but, still better, because it is to some degree indicative of the nature of evolution in the species consid- ered. ‘“‘Given any species in any region, the nearest related species is not likely to be found in the same region nor in a remote region, but in a neighboring district separated from the first by a barrier of some sort.’’ President Jordan has subsequently stated that “‘barrier’’ is here used in a very broad sense. This law gives the state of affairs that we should expect if evolution were mainly the work of isolation and were taking place by even flow rather than by waves or mutations. It is doubtless mainly his knowledge of the distribution of fish that led President Jordan to this formulation. Its applicability to the vertebrates is quite wide. Botanists do not seem to find it of as general applicability in plants; indeed, in some genera, like Draba, the opposite condition prevails. In invertebrates but few authors have considered it. It is important that many groups should be examined with this in view, for it throws valuable light on the method of evolution. We have seen that it is largely inappli- cable for the major species because of their extensive overlapping.

For the consideration of the distribution of varieties and minor species I have drawn the distribution of the several derivatives of Hippodamia convergens in fig. 91. The distribution of Hippodamia convergens itself is not shown. It is found thoughout the area of the map. It will be seen that Jordan’s law fails here, for the closest allied forms occupy the same region in all cases with the parent species and in many cases with an allied variety. Thus the closely allied variety quinquesignata and variety caseyi have ranges nearly coextensive in the west. So far as our data goes, variety caseyi does not share the eastward extension of variety quin- quesignata. It is characterized by the possession of three of the four dis- tinctive characteristics of the variety quinquesignata and is so generally associated with it that I believe future study in the eastern mountains will reveal its presence. Hippodamia bowditchi dwells in a part of the range of its nearest ally and is surrounded by it. The conclusion seems inevitable that we are dealing here with systematic units that have not arisen by an en masse evolution of all individuals in the locality, but have arisen at first in a few individuals which by virtue of some degree of seg- regation in heredity have escaped being swamped. This conclusion, reached from the study of the distribution, is the same as that which the study of the variation and heredity of these beetles leads to.

Geographical isolation does not act in these beetles as the initial cause of separation as it does with blending characteristics. It may assist the variety in its progress, however, (1) by preventing too wide outcrossing and thus causing the individuals of the variety to more frequently inter- breed, and (2) by helping the variety to attain intersterility and thus a specific status, in a way shown later. Specimens of Hippodamia spuria isolated from its close ally, Hippodamia interrogans, are more alike in gen-

GENERAL DISCUSSION. 91

eral in their color-pattern than are the extremes of Hippodamia spuria where there is no isolation.

Taylor (1894), Matthew (1908), and Cockerell (1908) have proposed the view of an “‘active evolutionary center’’ at which a genus or a higher group has reached its highest development and from which the new species arise and are dispersed. In this way the oldest species are found ata dis- tance from the center. Wedo have an active evolutionary center for Hip- podamia in the Plateau and Pacific States, but it is doubtful if the varieties of species there produced are dispersed and replaced. They are, for the most part, endemic and have no power to extend their range greatly into conditions quite different from those of their origin. Furthermore, the active evolutionary center is here caused by conditions of the environment capable of altering the germ-plasm rather than by any “‘intense pressure and competition there,’’ as Taylor asserts of the active evolutionary center of the land snails.

HEREDITY. In spite of the many non-conformable cases presently to be mentioned,

many biologists have come to expect Mendelian segregation and dominance whenever segregate inheritance is found, so eagerly does the mind wel-

- eome and embrace any conception of uniform action. In the discussion of

the several species we have seen that in only one case do the facts seem conformable where, as in several cases, they fail to conform to the simple Mendelian conceptions in spite of the fact of there being some degree of obvious segregation. For the biologist whose predilections are strongly Mendelian this is of litthe moment, for he will believe that these cases simply demand a few consistent hypotheses of additional factors, enzymes. inhibitors, coupling, latency, etc., to become quite conformable. Sincea rigorous test of most of these subsidiary hypotheses to Mendelism demand very large numbers, greater than are often available in animal experi- mentation, the test is not be made and consequently his comfortable faith remains undisturbed.

But is it good scientific method to keep protected in this way the faith in the Mendelian behavior of insect characteristics ? When we consider the results of Kellogg upon silkworms (1908), Lutz upon Ampelophila (1908 ¢) and upon Grioceris (1908 a), McCracken upon Lina (1907) and upon Gastroidea (1906), and Tower upon Leptinotarsa (1906), we find no genus in which there are not such un-Mendelian things as “‘ individual and strain idiosyncracy,’’ “‘variation of allelomorph potency,’’ ‘‘ weak factors,’’ “‘bilateral opposition of characters,’’ “‘ gradual elimination of alternative characters,’’ “‘ progressive dominance,’’ and “‘inability to hand on variations with full intensity.”’

The least we can do is to give a fair consideration to hypotheses which seem to arise simply from the facts, for, as Chamberlain has well pointed - out, the use of multiple hypotheses is a great safeguard to the soundness of scientific progress.

92 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

The hypothesis which seems to me to satisfy best the facts of heredity here and to be consistent with those of variation and distribution in these beetles is that of characteristic prepotency. Since the word ‘‘ prepotency ”’ is used in speaking of an individual, I would propose the term “‘ preponder- ance’’ for the prepotency of characteristics. It is necessary to distinguish this from dominance. In dominance without the aid of selection there is no gain in numbers, since there is no germinal advantage with the domi- nant characteristic. In preponderance there is a germinal advantage, so that the characteristic gains in numbers without the aid of natural selec- tion or even conceivably in opposition to it. In a character with blending inheritance, preponderance shows itself by the progeny resembling in the long run one of the two opposed characteristics more than the other.

In segregate characteristics it may show itself in the guise of hyper- dominance, that is, by the numbers of the dominant characteristic being excessive, or ina differential tainting of the extracted homozygotes. Cases that seem to be more easily explained by resorting to preponderance (or subponderance) are the deficient proportions of most of the elementary species of Oenothera where crossed with O. lamarckiana (de Vries, 1901) , the deficient proportions of abnormally veined Ampelophila when crossed with the normal (Lutz, 1907), the deficiency of polydactylism when crossed with the normal in man (Davenport, 1904), the excessive proportion of booting in booted < bootless in poultry (Davenport, 1906), deficient pro- portion of the all-black Lina scripta < normal (McCracken, 1907), exces- sive proportion of spotted Lina scripta in L. scripta black L. seripta, and excessive proportion of black Gastroidea dissimilis in black < green (McCracken, 1906).

Many breeds of animals have what are called “‘faults,’’ which are the distress of their breeders. These faults are sometimes merely the normal traits of the wild species or ancestral variety from which the variety is obtained, and their persistence would not be a serious problem except that the breeder is trying to handle so many characteristics. There is a greater degree of persistence in some, however, that can not be thus explained. Of recent years, these have been explained as the products of the meeting of heterozygotes which reveal recessive characters which may not have been seen for several generations, because the heterozygotous individuals had not mated before. This is sometimes true without doubt, but in con- sidering these cases from Leighton (1907, p. 321), and Simpson (1903, p. 117), the hypothesis of preponderance may well be considered. Here the preponderance shows by the difficulty with which the characteristic is submerged.

‘Litters [of the white English terrier] frequently show the blemish of a spot of brindle or russet. These spots usually appear behind the ears or on the neck and are of course a disfigurement on a dog whose coat to be

* perfect should be white.’’ This breed was quite popular at least 40 years

ago, and any color or marking disqualifies.

GENERAL DISCUSSION. 93

“‘In most black litters [of black Persian cats] at least one has a white spot on chest or throat, and this generally settles the question which, if any, shall join the majority at a tender age. The unfortunate kit’s pedi- igree may be absolutely devoid of offense on this point. Apparently no precautions can prevent or eradicate the fault.”’

From the instability of some characteristics we may infer that they would be subponderant in heredity. Thus in Leighton, page 403, refer- ring to the Dandie Dinmont terriers, we find the following: “Tt is gener- ally best to mate a mustard to a pepper to prevent the mustards becoming too light in color, though two rich-colored mustards may be mated with good results.”’

On page 87, ‘‘ When brindle Great Danes are continuously bred together it has been found that they get darker and that the peculiar striping dis- appears and in that case the introduction of a good fawn into the strain is advisable. The constant mating of harlequins (black spots on white) has the tendency to make the black patches disappear and the union with a good black Great Dane will prevent the loss of color.’’

The tendency of most white breeds of poultry to become cream-colored and of Barred Plymouth Rocks to become “‘smoky”’ asserts itself when- ever selection of a good strain ceases. The stalwart Mendelian, in this - Jast case, has an explanation in the assumption that ‘“‘smoky’’ Barred is a unit-character, and that light Barred is only produced by the selection of fluctuations and hence must inevitably fall back when selection ceases. But this is nothing more or less than the assumption of the preponderance of the type of the unit-character over its derivative. The stalwart Men- delian then admits preponderance in fluctuations so long as they lead to a unit-character. If, then, there is nothing intrinsically impossible about preponderance, the only objection to it as a factor of evolution to the Men- delian lies in his conception of clear-cut unit-characters having a perfect segregation. Any belief in such sharply-cut phenomena is untenable in the consideration of the color-pattern of these beetles, where some degree of segregation is nevertheless positive.

Cases of mass reversion are, to my mind, often the result of preponder- ance of the reverting character over the newer conditions. Of course reversion taking place in cases following the cessation of short, active selection is to be attributed to regression.

The degree of segregation in heredity is most perfect in spots 1 — 3 in Hippodamia convergens. It varies by easy stages to cases in which the heredity is nearly if not quite blending, as in the relation of 5 and 6 in Epilachna borealis or as in the incomplete pronotal margin or in the size of the pronotal dash in Hippodamia convergens. Just short of blended inher- itance we have such cases as those in which the bulk of the individuals lie between two extremes, at each of which many individuals are massed. The result may be likened to the balls on a bowling alley, which, while passing frequently the position of each pin, yet in the long run pass more

94 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

frequently into the side-gutters than any other one place. Only a sugges- tion of segregation is seen in 5 and 6 in Epilachna borealis. Here we seea variation curve strongly-skew in the direction of confluence of spots, and the experiments in heredity give us progeny which seem for the most part intermediate, but nearly always there are many without confluence and some with fully as much or more than the parent. Here, then, confluence is not swamped, because, in spite of the majority of intermediate and normal progeny, a mechanism, in some ways suggestive of that which produces de Vries’s ever-sporting varieties, maintains the supply of these confluent individuals.

In Coccinella novemnotata we have a type of heredity that deserves fur- ther mention. Here the variation consists of a series of steps. We have 4 spots in which we may have no confluence, or confluence of 2, of 3, or of 4 spots. Intergrades exist in which there is an approach to confluence or partial confluence; nevertheless the steps referred to constitute centers of variation more frequented than intermediate conditions. Now, the progeny may belong to the centers of variation of their parents, but also to the other centers. The variation is, then, moniliform and the heredity poly- morphic.

The facts of variation and distribution in these genera show us that the swamping influence of intercrossing is not effective upon a large number of varieties. This immunity is obtained by some degree of segregate heredity. Absolute segregate heredity is not found, hence we may con- clude that a variety may be protected from swamping by an only partly segregate heredity.

PHYLOGENY.

It is desirable if possible to decide upon the primitive pattern from which the coccinellid beetles were descended. Schoeder, while admitting that the pattern of Adalia bipunctata and its varieties are referable to a scheme of 7 spots and a scutellar mark, contends that the primitive pattern was one of vittz. This was followed by barring, and a reduction of the network thus formed left the points of intersection as spots.

That the primitive pattern was one of vittee may be possible if we go back to a sufficiently remote time. It is possible that coccinellids having as regular vittez as those of Paranaemia vittigera have retained this pat- tern, with possibly some reduction in the number of vittee. Tower (1906) finds that vittze in Leptinotarsa are between the veins, yet the extension of pigment along the veins in modification experiments in the lady-beetles makes it possible that the vittze were here differently placed. In Adalia these spots are upon the veins, but this can not be said to be the general rule. The position of the spots seems to be remarkably independent of the veins. The influence of the vein is much more evident in projections from the spots along the linea externa in Coccinella novemnotata and Harmonia

picta.

GENERAL DISCUSSION. 95

The term ‘‘primitive’’ is a relative one and, in the working out of the evolution of the color-pattern, much more value lies in the ascertainment of the ancestor of the genus or subgenus than of a more remote ancestor. There are not enough data at present and there probably never will be to make the attempt to study the evolution of the color-pattern from genus to genus profitable, and this is not attempted in this paper. The study of evolution is served better by concentration upon the evolution of the pat- tern within the genus, subgenus, or species, and this has been my task. The distribution and number of spots differ in the different genera. These differences I believe have arisen by the different ways in which the vittz have given way to spots in the phylogeny of the genus.

The beetles of each genus, with a few exceptions, where the unit is a subgenus, are referable to one set of spots. They constitute a pattern- unit. Within this pattern-unit the work of evolution is principally the loss or confluence of spots. Only rarely are new spots added. Change or shape of elytron sometimes shifts the position. Otherwise the position is very conservative.

The reasons for believing the spotted pattern to be primitive for Hippo- damia at least are (1) the wide distribution of the spotted species, and (2) the narrow distribution and varietal nature of so many of the forms which have deviated from the spotted condition. But so difficult is the change of number and position of spots between the several genera that it is easier to conceive of them arising by independent origins from some primitive vittate condition.

EVOLUTION.

We have seen, in the preliminary discussions, that natural selection must be very feeble in the evolution of the color-pattern of coccinellid beetles. Since the pattern is for the purpose of association with the bad taste, if it has a purpose, its highest utility would lie in constancy and idiosyncrasy. The spotted pattern being in these beetles the commonest and character- istic one, natural selection, in so far as it is operative, should favor this pattern. If, then, we find any evolution away from this pattern, it must have taken place either without the aid of natural selection or in opposi- tion to it. There is occasion for some doubt as to the primitive pattern of some of the genera and species. Let us consider, then, a case where the evidence that the spotted pattern is ancestral is conclusive, viz, Hippo- damia convergens and its varieties. Here we have a species with the same Spotted pattern that is possessed by several other species of the genus and which is widely distributed. In this beetle the spotted condition gives way to marked deviations from its original spotted condition in several ways in different varieties (fig. 11).

The question is, What has caused this*evolution? Recourse to natural selection is debarred. The inheritance of somatogenic characters is open to such grave questions about the conceivable mechanism as to be very

96 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

doubtful. It is clearly inadequate to produce characters like the extensa mark or the spot g, which environment could not produce acting upon the soma alone. Further, many of the somatogenic characters do not become germinal, as the dark linea externa in Hippodamia convergens. On the other hand, we have seen, in studying the variation and heredity of the markings, evidence of determinate variation and of preponderance. It is to this first factor, aided, I believe, by the second, that this evolution is to be attributed. Since their action is analogous, I have grouped them under the name of determinate evolution in the title of this paper.

Determinate variation is a dual process, for although the environment frequently causes certain variations, as will be shown later, the nature of the germ-plasm is a more important element in determining whether there shall be any variation formed and what it shall be. In other cases the environmental control of the germinal change is so minor a feature that it is negligible and the cause may be said to lie in the nature of the germ- plasm. The relative réles of environment and the constitution of the germ-plasm are interesting in the case of disappearance of spots, where we found in one environment one spot especially disposed to disappear- ance, whereas in another environment it is another spot (fig. 17). The role of the environment is shown in the following cases: On the Pacific coast, outside of the mountains and the humid coast region, there is a strong tendency to disappearance of spots, five species yielding to this tendency. In the mountains, transverse confluence reigns. Inthe humid Pacific coast region Hippodamia convergens gives us the very dark variety moesta. It is possible that marshes evoke vittation, for Hippodamia sin- uata (trivittata) is a marsh-dweller, as is Paranaemia vittigera, while Naemia seriata is maritime; but this must be further investigated. Arid regions lead to the reduction of spots, although the greatest proportion of spotless individuals is not found in the region of greatest aridity.

Where two species are affected by one of these environments, we should get corresponding results. This is realized in the Cascade Mountains, where we have the barred Hippodamia oregonensis, a derivative of H. spuria, and the barred variety quinquesignata from H. convergens. In the Hudson Bay region, with the vittate H. falcigera we get the vittate deriv- ative of H. tredecimpunctata—the Hippodamia americana.

In the Pacific coast region outside of the mountains and in the humid region, spotlessness is found in Hippodamia convergens obsoleta, H. tre- decimpunctata var. immaculata, Coccinella perplexa var. subversa, Coc- cinella novemnotata var. franciscana, and Coccinella californica, probably a spotless derivative of Coccinella transversoguttata. Yet it should be mentioned that Hippodamia parenthesis, apicalis, and spuria resist the general trend towards spotlessness.

The role of the environment must not be overemphasized, for its part may be likened figuratively to the heat that starts a chemical reaction.

“Where heat is applied, the result differs with the different chemicals

GENERAL DISCUSSION. 97

affected. No reaction may result, comparable ones may result, or wholly different ones may result, and besides, different degrees of the heat may be necessary to give any of these results.

Adalia bipunctata in Europe has many varieties, modified in the direction of greater pigment, and many of them are found in any one locality (Weis- ner, 1907), yet in America this species is relatively constant. Coccinella perplexa and its very close European ally, C. trifasciata, with an elytral pattern practically identical, are quite constant, except in the Pacific States, where there are many varieties of reduction. Hippodamia par- enthesis maintains a greater constancy in the Western States, where so many species become especially prolific in varieties than it does in the Eastern States.

In the Eastern States most of the species do not find the conditions such as to set up new variation lines. Yet Coccinella novemnotata gives us its varieties of vittation in the East, and throughout the Plateau and Pacific States it offers no varieties but those of reduction. Megilla macu- lata and Naemia seriata suffer reduction of spots in the Southwestern States, where the other species are also thus affected. Hippodamia api- calis suffers a greater reduction in Eastern Colorado than in the Pacific region.

Determinate variation governs the birth of a variety, but its later career is governed largely by its method of inheritance and by natural selection. If the variety is subponderant, it will probably take the status _of a rarity, arising again and again by the determinate variation but never winning an important place. It is possible that subponderance in some characters is capable of being converted into preponderance by selec- tion or by the action of environment on the germ-plasm and a place thus acquired by it. In others their subponderance may be expected to be fun- damental and unalterable. If the characteristic be preponderant, its way is made smoother, and if determinate variation cooperates by frequently offering the variation, the characteristic is sure to make its way unless opposed strongly by natural selection. If it arises very rarely, its fate is in jeopardy. A mutation which arises only once has a very small chance of success. The frequency with which it arises is an important element in its fate.

This analysis is carried further in table 21. We see, in examining this table, why a variety aided by determinate variation does not always sup- plant the species, but often reaches a status of a certain degree of relative abundance which it may retain for a long time. Thus in Oregon the vari- ety of Hippodamia convergens with open pronotum continues to exist as a rarity, avoiding extermination on the one hand and increase on the other, because it is kept in existence by determinate variation, and, as I believe, is kept repressed by its subponderance. Since the activity of natural selection, determinate variation, and preponderance are each found in

98 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

various degrees, the categories used in the table are of course not sharply limited, but merely arbitrary points taken for exemplification.

There can be no one general orthogenetic influence here, so diverse are the many varieties which have arisen from Hippodamia convergens in dif- ferent regions. Eimer’s law, that a primitive pattern of stripe is followed by spots, which in turn give way to tranverse striping, and last of all to a uniform color without spots or stripes, can not hold. In fig. 90 the vari- eties of Hippodamia convergens are so arranged as to show how one degree of one line of development may be combined with another degree of another lineof development. In fig. 29, in Hippodamia spuria, a similar condition is met with, even though here we deal with both stripe and bars. Another law of Eimer’s, “‘new markings appear upon the body of an animal from behind forwards and above downwards, or conversely, whilst the old ones disappear in the same direction and succession,’’ isalso untenable. The study made of the order of disappearance of spots does not bear out this rule, nor does the appearance of new spots, suchas qand ein Hippodamia convergens, conform toit. The diversity of patterns here makes impossible the formulation of any general orthogenetic law, for, as we have seen, they progress differently from place to place. In fact, the general diver- sity of the animal kingdom throws great doubt on all general laws of orthogenesis. The doctrine of determinate variation is only handicapped by the formulation of these rules, which must be so soon abandoned. Determinate variation is as multifarious as are the phenomena it aims to explain.

TABLE 21.—Effects of cooperation or of opposition of some evolutionary factors upon a new characteristic.

| Heredity Heredity Heredity subponderant. indifferent. preponderant. Natural selection— i i ; Wat : hear one Variations | basses a aos Doi area as arise arise : arise selatively. | relatively | Jelatively | relatively | TORNVCY, | relatively quently. frequently. quently. frequently. quently. frequently. oie Scents ODOBES. «oc so sca ne 8's Remains Remains Remains | Remains Doubtful Doubtful extremely rare. | very rare. rare. result, result. rare. | | Indifferent,.......... Remains | Remains | Remains | Remains May Increases. very rare, rare, rare. uncommon.}| increase. PAVOEN sesh cose as aes Doubtful Doubtful | May Increases, _ May Increases result. | result, increase. increase very | rapidly. rapidly.

Evolution may conceivably proceed in a regular and continuous fashion, and it has been generally so conceived, until of recent years, when much attention has been paid to the evolution by leaps. Considerable evidence, as to positions of organic stability and unit-characters, is held to substan-

tiate this view. In these beetles we would seem to have both saltatory evolution in a character like 1 + 3 in Hippodamia convergens and evolu-

+ > oe ‘

—-

SY dik

3

GENERAL DISCUSSION.

3 6 oh OS Fig. 92.—Hypothetical polygons to illustrate centers of variation and evolution by waves.

100 EVOLUTION IN COLOR-PATTERN OF THE LADY-BEETLES.

tion by even flow in the lateral process of the black area in the pronotum. But, in addition to these, I believe we have a third intermediate method, evolution by waves; for most of the characters, while showing some degree of discontinuity in variation and segregation in heredity, are not so dis- crete as to be properly called unit-characters. They are better described by Galton’s expression ‘“‘positions of organic stability,’’ interpreting position in a broad sense. I prefer the expression “‘centers of variation.’’ Thus, in the size of spot 1, instead of two unit-characters at 5 and 15 units diameter,it seems to me we have more probably two centers of vari- ation at these two points. This distinction would be of little importance were it not for some important evolutionary consequences.

The centers of variation (and here I am not speaking of the clear-cut cases of unit-characters) may be the result of either of two possible causes. In one kind, the germinal centers of variation, the positions of organic stability are those of the germ-plasm, depending upon the nature of its structure and processes. In the other kind, that of somatic centers of variation, we have positions of organic stability of the soma, the germ- plasm in this case not showing any corresponding favor for one degree rather than another. We may illustrate this latter type by considering the evolution of a pattern from one in which two spots are separate to one in which they are confluent. It seems to me not only possible but probable that it is easier for the spots to develop separately or in full confluence than with a narrrow connecting band. Now, as the determiners move on in the direction of confluence (whatever may cause the movement), indi- viduals which have intermediate determiners will be more likely to have the confluence either less or more than that determined by the germ-plasm. The result will be a bimodal polygon of frequency, until the germ-plasm has progressed far enough to carry all the individuals past the position of disfavor.

In fig. 92 I have aimed to illustrate this in a hypothetical case. For the polygon of frequency I have taken the one used in Davenport (1904) to illustrate the normal curve. Now, I have assumed arbitrarily that the magnitude 20 of the character is a position of organic stability, being favored at the expense of other magnitudes to the extent that it receives eight times as many individuals as it would otherwise receive; that the neighboring classes are favored fourfold; those next adjoining twofold; those next are unaffected. The next three classes are reduced to the extent of getting only one-half, one-quarter, and one-eighth, respectively, of their quota. The numbers thus obtained have been reduced to percent- ages. The polygon of frequency centered over 20 with the changes thus produced is shown as the first position in table 22 and fig. 92. Now, let the character evolve by the increase of the magnitude by any factor of evolution. The polygon will move to the right and assume the succes- sive positions shown in table 22. In fig. 92 these curves are graphically represented, with a few omitted to avoid overcrowding.

GENERAL DISCUSSION. 101

The polygon can not move as an undisturbed whole, but will start the formation of a new polygon at the next center of variation. With the increase in the polygon at the new center, the old polygon decreases. Such a phenomenon is generally thought to be necessarily that of a unit-char- acter, but we have seen that no such germinal discontinuity is necessary to account for it. Theterm “‘unit-character’’ should be reserved for those cases of truly distinct units analogous to chemical compounds or to the rungs of a ladder, rather than to the dentation or crenation on a leaf- margin. But, if we grant the existence of such centers of variation as figured, then, in addition to evolution proceeding by even fiow and by leaps, it may also proceed by waves or undulations.

TABLE oS Lae age case of somatic centers of variation.

Normal fre 2| 6 10 18/82 48 64 74 | 80 | 74

| 2113) 14) 5 6 7 18 19 2 21 2 23 24| 25 26| 27 64 48 ‘teas Ses aaa 6

4 | |

Successive positions of the shifting polygon

heart 1} .1. 1) .2/1.0'2 9) 7.8118.139.218.17.82.91.0) .2

Eo P| al

SS SRNR

;

'

|

| ' cd

mary on St 4 Al 3) .3) 9 2.3(5.3)11 5 21.3/36.9 13.84. 61.3 .110.1

5 é. :

' | I. | --|X ax BX ax HX aX 1x 2} x*4)x 8x aps ax Lx sarap aix 4x 8) x 4hc2) x1] 5h se ) oe 2 ai ae 0 Ed oe ee 2a 1-10.1-| 0.1, 0.2) 1.0) 2.9 7.818.139.218.1 7.8 2.9)1.00.20.1/0.1)0.1 || | i) -6) 2.1) 6.216.437 9.20.5) 9.5 4.1/1.5 5.1) .1) .210.1 ae 1.8 4.618.825.9213. 5] 5.32.8 9 3} 8) .4f 4} O.s) || | S| .8| $.111.193.3/22.212.8 6.938.214 5 7) 8) 8} 1.0 07 a a} 722 7 $27.920.914.1 8.1/4.8 2.0 .9|1.3|1.7/2.0] 2.2 2.6) 1.7 316 5.3 19.217.012.8 6.5/4.92.71.211.7/8.24.3 4.8; 5.3) 6.4 1.1 6) 3.511 a 9. 37.0 7.0|4.62.6 1. 26 6/4.67.0 9.3/10.411.6 3.50.6 pq 11 64 53/48 4.518.217 1.22. 74. 9/8.5|12.817.019.2 5.31.60.3 113/26 221201 71.3 92.05. 3/8.1/14.120.927.9 7.82.2 .70.1 7| 10 8 8 7 Bi. 43. 2/6.9|12.822.933.311.131 .8 .3

2

tee 2 1 4 5154.1) 9.5/20.5'87.9 16.46.22.1 -6) 201 -10.1-'0.1-

Xt

SA

| } |

'

25) 29) a0 31 | 32 | 33 | a4| 35| 26, 37, 38 39 | 40

Number reduced to percentages.-

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 convergens 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 somatogenie 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 isin 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) Phylogeny.—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 thespotted 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 by mutation in different characteristics at different times.

BIBLIOGRAPHY.

Burcess, A. F. 1898. An any ny Coccinellid. U.S. Dept. of Agriculture, Div. of Ento. Bull 17, 59-61. 1903. Moseminiie Pastas on the family Coccinellide. 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, Passer domesticus. Biological lectures delivered at the Marine Biological Labora- tory at Wood’s Holl, in the summer sessions of 1897 and 1898, pp. 209-226. Crotcu, G. R. 1873. Revision of the Coccinellide of the United States. Trans. Am. Ent. Soc., : vol. 4, pp. 363-382. 1874. A Levey of the Coleopterous family Coccinellidz, London. 1898. The zonal distribution of Coleoptera. Bull. 28, Agr. Exp. Sta., Mesilla Park,

N. Mex 1908. Fossil Aphiitidles from Florissant, Colo. Nature, vol. 78, pp. 318-319. Casey, T. L. 1899. A revision of the American Coccinellide. Jour. N. Y. Ent. Soc., vol. vit,

70-163. CRAMPTON, H HE.

1904. Experimental and statistical studies upon Lepidoptera. I. Variation and elimination in Philosamia cynthia. Biometrika, vol. 3, pp. 113-130. DAVENPORT, C. B. 1904. Wonder-horses and Mendelism. Science, XIx., 151-153. 1904. Statistical methods with special reference to Rolagica! variation. Second edition. New York. 1906. Inheritance in poultry. Carnegie Institution of Washington Publication No. 52. 1908. faerie in canaries. Carnegie Institution of Washington Publication ° DE VRIES, H. 1901-1903. Die Mutationstheorie. Leipzig. DimMock, G. W. 1906. Algunas Coccinellidze de Cuba. Primer Informe Anual de la Estacion Central rm Agronomica de Cuba, pp. 287-392. FitTcu,

1861. Sixth report on the noxious, beneficial, and other insects of the State of New York. Albany.

F. 1908. Wild Jamaica cotton. Science, xxvii, 307-308. GULICK, J. T. 1905. Evolution, racial and habitudinal. Carnegie Institution of Washington Pub- lication No. 25 Harpy, G. H. ee Ste sma proportions in a mixed population. Science, n. s., vol. 28, pp. 49. D 1899. Efficiency of some protective adaptations in securing insects from birds. Am. Nat., vol. 33, 461-484. JOHNSON, R. H. 1907. me ta on aphids and coccinellids. Ento. News, vol. XVIIL., pp. KELLocG, V. L., and R. G. BELL. 1904. Studies of variation in insects. Proc. Wash. Acad. Sci., vol. vi., pp. 203-332. , VERNON L. 1908.” Inheritance in silkworms, I. Leland Stanford Junior Univ. Publication. Uni- versity Series No. 1. LENG, C. W.

1903a. Notes on Coccinellidze. I. Hippodamia. Journ. N. Y. Ent. Soc., vol. XI, pp.

35-47.

19036. Notes on Coccinellidz. II. Coccinellini. Journ. N. Y. Ent. Soc., vol. x1, pp. 193-213.

1908. Notes on geet Il. Chilocorini. Journ. N. Y. Ent. Soc., vol. xvi, or 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 Crioceris asparagi. Psyche, vol. xv., pp. 50-52.

1908c. Tage of alternative and blending inheritance. Science, 28, pp.

MEISSNER, O.

1907. ivie relative Haufigkeit der varietaten von Adalia bipunctata 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 bipunctata L. in Potsdam (1907), nebst molgplenese Bemerkungen iiber diese und einige andere Coccinelliden. Z. f. wiss. Insektenbiologie, Band III., pp. 309-313, 334—

MULSANT, E. 1846. sag eh Naturelle des Coleopteres de France. Sulcicolles et Securipalpes. aris. 1850. Species des Coleopteres trimeres’securipalpes. Paris. 1866. Monographie des Coccinellides. 1° Partre. Coccinelliens. Paris. MATTHEWS, W. D. 1908. Osteology and Phylogeny of the American Cetvide. Bull. Am. Mus. of Nat. Hist., vol. XxIv, pp. 535-562. MacDouaGa., D. T., A. M. VarIL, and G. H. SHULL. : 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 Amblystoma tigrinum. University studies. Pub. by Univ. of Neb., vol. vil, pp. 197. REICHERT, ALEX. 1896. Uber Coccinelliden und ihre Variation. Illus. Wochensch. Entomologie, Band L., pp. 26-30. 1904. Die Varietaten von Adalia bipunctata. Entom. Jahrb. fiir 1904, pp. 179-181. RATHVEN, A. G. : 1908. Mice en genetic relationships of the garter snakes. U.S. Nat. Mus. ull. 61. SIMPSON, FRANCES. 1903. The book of the cat. London. SHULL, 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- tarsa. Carnegie Institution of Washington Publication No. 48. WEISE, J. 1879. Bestimmings-Tabellen der europaischen Coleoptera. II. Coccinellidz. 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. 1908. Notes on the Coccinellide. Can. Ento., xL., 393-421. November, 1908.

7 Wy

:

i. 7

me Cera ape HO

ae - e

oe a

LW ANS Ms

2 ew c i: = ag Se fe a Wa, OS a ag - < zo OF wy M 4 : > Sn ra Se 8 fe

. + ey. JO Ure, ed-TOTOS Sy UT WOT4YNTOAS Sq eUTTSZ Sq

*88T100q~ApBT

: By 09SUT TeoZ

TITH TTemsoy “uostiqor sony

eS TQOT

Sea tities tea

es

ety thoy mag ates ieee race rastarin tin x ‘ * eters Sok x) * ata % : : ie tei tp ie = the Hee aseht tr apnea retary sade. 7 4 NY A 5 . " st ise r it) SG? ese

*

x von He