In all discussions of the modes of Evolution the phenomena of Geographical Distribution have been admitted to be of paramount importance. First came the broad question, were the facts of distribution consistent with the Doctrine of Descent? I suppose all naturalists are now agreed that they are thus consistent, and that though some very curious and as yet inexplicable cases remain to be accounted for, the distribution of animal and plant life on the face of the earth is much what we might expect as a result of a process of descent with modification. Passing from this general admission to the more particular question whether the facts of distribution favour one special conception of the mode of progress of evolution rather than another, no agreement has yet been reached. One outstanding feature is hardly in dispute, namely that prolonged isolation is generally followed by greater or less change in the population isolated. Groups of individuals which from various causes are debarred from free intermixture with other groups almost always exhibit peculiarities, but on the other hand, cosmopolitan types which range over wide areas are on the whole uniform, or nearly so throughout their distribution. Examples of these two categories will be familiar to all naturalists. The barriers to intercourse may be seas, deserts, prairies, mountain-chains, or circumstances of a much less obvious character which isolate quite as effectually. The local unit is not necessarily an island, a district, or an area of special geological formation, but may, as every collector knows, be a valley, a pond, a creek, a "bank" in the sea, a clump of trees, a group of rocks in a bay, or a particular patch of ground on a mountain side. All the great groups provide examples of such specially isolated forms. The botanist knows them well; the conchologist, the entomologist, the ornithologist and the student of marine life are all equally aware that special varieties or special species come from special places and from nowhere else. In one remarkable case the season of appearance plainly acts as the isolating barrier. Tephrosia bistortata is a small Geometrid moth which has two broods, appearing in March and July respectively. It is closely allied to T. crepuscularia which emerges in May and June. From the fact that occasional specimens cannot be quite certainly referred to one or other of the two, many have held that the two are one species. Nevertheless, in general they present distinctions which are plain enough. Some localities have one form only, but in several woods they co-exist. Experiment has shown that the two can be crossed, and that the cross-breds can breed inter se and with at least one of the parent stocks.[1] Some diminution in fertility was observed, but perhaps not more than is commonly encountered when wild forms are bred in captivity. In such a case it can scarcely be doubted that the distinctness of the two forms in the places where they co-exist is maintained by the seasonal isolation.

Just as the consequences of isolation are to be seen in the most different forms of life so may they also affect the most diverse features of organisation, such as size, colour, sculpture, shape, or number of parts. In the Sloth (Choloepus) the geographical races differ in the number of cervical vertebrae—or in other words, in the distribution of vertebral differentiation. The geographical races of Cistudo differ in the number of claws and phalanges.[2]

In Shetland, the males of Hepialus humuli (the Ghost Moth) are not sharply differentiated in colour from the females, as they are elsewhere, but in varying degrees resemble them.[3] No such males are found in other localities, and even in the other Scottish islands they are normal. In the island of Waigiu the converse phenomenon has been observed in Phalanger maculatus. Generally the male is spotted with white, and the female is without spots, but in Waigiu the females are spotted like the males.[4]

The following striking illustration was pointed out to me by Dr. W. D. Miller. Euphonia elegantissima as it occurs in Mexico and Central America has the two sexes very distinct from each other. The male has the lower parts orange and the upper parts a dark indigo blue, with a bright turquoise-blue head and neck. The female, except for the head, is of a bright olive green. A form in which the sexes are similarly differentiated exists in Porto Rico and is known as E. Sclateri. But in many of the other West Indian islands the representative "species" (E. flavifrons) has the two sexes closely resembling the female of E. elegantissima. This form is found in Antigua, Barbados, St. Vincent, and Guadeloupe, from which localities the British Museum has specimens. All three so-called species are very much alike otherwise.

In the genus Pyrrhulagra (Loxigilla) to which Mr. Outram Bangs called my attention, several distinct and alternative possibilities occur. The genus has many local species occurring on the various West Indian islands. These species are characterized by differences in size, colour, and the shape of the bill. The colours have a narrow range, being black or greyish, with or without chestnut marks about the head and throat. In most of the islands the males are in general colour a full black, and the females are distinctly grey. They are thus found in San Domingo, Jamaica, Bahama, and most of the Lesser Antilles. In Porto Rico we meet the peculiarity that the hens are almost as black as the males (Ridgway describes the black of the hens as slightly less intense). This form is called portoricensis. A larger type, known as grandis, similarly coloured, inhabits St. Kitt's. Then, on the contrary, in Barbados, both sexes are a dull blackish grey, like the hens of the Lesser Antilles in general.

The local species of Agelaius show similarly capricious distinctions. A. phoeniceus is a widely spread species, found over a great part of North America. The male is black with red-orange bars on the wings, but the female is somewhat thrush-like in colour. In the island of Porto Rico there is a form called xanthomus, in which both sexes are like the males of the mainland. A similar species called humeralis, also with both sexes male-like, lives in Cuba. The island of Cuba, curiously enough, has also a distinct species named assimilis, in which the female is a dull black all over, though the male is like the mainland type.

So also may local races differ in respect of variability. Argynnis paphia, the Silver Washed Fritillary, through a great part of its distribution has only one female form. In the English New Forest a second female form, valesina, co-exists with the ordinary paphia female. But in the southern valleys of the Alps the valesina female is much the commoner of the two, and indeed in some localities where the species is abundant, I have seen no paphia females in many days collecting.

The beetle Gonioctena variabilis furnishes an illustration of a comparable phenomenon affecting the male sex. In 1894 and 1895 I studied the curious colour variations of this species especially in the neighbourhood of Granada, and Mr. Doncaster ten years later repeated the observations on the same ground, and also collected the insect in other places in the south of Spain. The distinctions are not easy to give in words and the reader is referred to the colour plate accompanying my paper.[5] The essential fact is that the males commonly have the elytra red with black spots and the females for the most part have greenish grey elytra with black stripes. In some localities a large minority of males closely resemble the female type, being identical in colour and then only distinguishable by structural differences. In two Granada localities I found the proportion of such males quite different. In the Darro valley about 38 per cent. (in 718) were of this feminine type, but on the hills some 300 feet above only 19 per cent. (in 3,230) were like the females. At Castillejo, not far from Toledo I found no such male in 75 specimens.

Mr. Doncaster collected from several localities, especially from two areas near Malaga, about 5 miles apart. In one of these the female-like males were, as usual, in a minority, but in the other these were actually in great excess, amounting to about 81 per cent. in the 173 taken. Doncaster found a doubtful indication that the composition of the population varies with the season, which is quite possible, but it is most interesting to note that in my chief locality after the lapse of ten years he found the proportions very much the same as I had done at the same season, for where I had 19 per cent. of the female-like males his collecting gave 16 per cent. In other respects also, his statistics corresponded very closely with mine.[6]

The various forms of Heliconius erato are well known to entomologists. They are strikingly distinguished by the colours of the strong comb-like marking on the hind wing, which may be red, yellow, green or blue. In various parts of the distribution in South America sometimes two and sometimes three of these distinct types co-exist.[7]

The distribution of the varieties of Noctua castanea typifies a large range of cases. The form which is reckoned the normal of the species has red fore-wings. It is practically restricted to Great Britain and Germany, according to Tutt. The other common form, neglecta, has grey fore-wings, and in this pattern it ranges through West Central Europe from North Italy to Germany. In the British Isles it extends up to Orkney. In Britain this grey form is by far the commoner, occurring wherever the species is found. The red form is much scarcer in England, and does not occur at all in many localities where the grey form is common. Mr. Woodforde, from whom this account is taken,[8] states that in August, 1899, he saw considerably over a hundred of the grey in the New Forest at sugar, but only two red ones. In Staffordshire however the red is proportionately more numerous and he estimates them as 40 per cent. of the population. Lastly a form has been taken in Staffordshire as a rarity in which the red is replaced by yellow, and this has hitherto been seen nowhere else. It is beyond our immediate purposes to discuss the genetic relationships of such forms, but the details of this case are interesting as making fairly clear the fact that the distinctions between castanea and neglecta are due to combinations of the presence of and absence of two pairs of factors, of which one produces a red pigment in the ground colour of the forewing and the other irrorates the same region with black scales. Mr. Woodforde states that all intermediates exist, and that in Staffordshire the greys always have a pinkish tinge. The yellow is doubtless another recessive to the red.

Species which are uniform in some localities may be polymorphic in others. Such a phenomenon is well exemplified by the orchid Aceras hircina. Of this species distinct varieties had previously been known in Germany, but GallÉ[9] has lately given a detailed account of a number of most diverse forms found growing in a district of Eastern France. Without reference to his plates it is impossible to give any adequate conception of the profusion of types which the flowers of the species there assume. In some the lip is elongated to many times its usual length, twisting and dividing in a fashion suggesting some of the strangest of the Tropical Orchids. In others the labellum and the lateral petals are all comparatively short and wide (Fig. 13). Intermediates, combining these qualities in various degrees, were abundant, and the condition of the species, which was the only representative of the genus in the locality, recalls the extreme polymorphism of many of the Noctuid Moths.

Fig. 13. Various forms of Aceras hircina. (After GallÉ.) This figure only shows a few of the more striking forms illustrated in GallÉ's plates.

Somewhat comparable variability has been seen in another Orchid genus Ophrys. In Great Britain the species apifera, aranifera and muscifera though variable are fairly distinct, but Moggridge has published two series of plates[10] showing a very different state of things as regards the Ophrys population of the Riviera. Here the outward diversity is such that the ordinary specific names cannot be applied with any confidence and the limits of the species are quite uncertain. It may well be supposed that these Riviera plants are interbreeding, and indeed we may safely assume that they are. It is, however, to be remembered that Darwin showed apifera in this country to be habitually self-fertilised, so that the different behaviour on the Riviera may itself constitute a local peculiarity. Moreover it is to be gathered from Moggridge's account that in the districts which he examined the condition was not to be described by the statement that our three types were there co-existing and hybridising, but rather we should say that the population was polymorphic, containing these three types amongst others. Conchologists are aware that on the Dogger Bank Modiola attains a size unparalleled elsewhere. The same is true of the sponges Grantia compressa and Grantia ciliata in the estuary of the Orwell.[11] Conversely, as we know so well in the case of Man, dwarf races occur in several special localities. Such examples may be multiplied indefinitely.

The relation of local forms to species has often been discussed from many points of view, but I know no treatment of the subject clearer or more comprehensive than an excellent account of some of the various manifestations of local differentiation as they appear in HelicidÆ published by Coutagne[12] and a reader interested in the problem which they raise would do well to make himself acquainted with the original from which the following notes are taken. He speaks for example of Helix lapicida. This is on the whole a constant form ranging up to the altitude of 1,300 m., common all over France except at great heights and in the Olive regions where it is restricted to moist places. Though subjected to such diverse conditions it shows only trivial variations in colour and other respects throughout its distribution, excepting that on both sides of the Pyrenees it has a very distinct sporadic variety called Andorrica or microporus. This variety occurs here and there, together with the type-form sometimes in colonies (pp. 26-30 and 86).

Bulimus detritus though more restricted in geographical range is a much more variable form. It exhibits great variations in colour, form, and size, and as Coutagne well insists, these are independent of each other. Foreshadowing the methods of factorial analysis he suggests that distinctions in each respect, the "modes" as he calls them, should be denoted by a letter, or if desired, by a name, and the several combinations of differences might thus be most logically and usefully expressed. Of such combinations he says there are at least 18, all of which can be found. The whole possible series does not necessarily occur in the same place, and various localities are characterised by the presence or absence of certain of the combinations as Coutagne calls them, and by the relative frequency with which they occur. The ideas thus enunciated are much in advance of the ordinary practice of systematists, who give names to forms which are nothing but accidental combinations of factors, just as the horticulturists for practical reasons give names to similar combinations, which as we now know are merely specially noticeable terms in a long series of possibilities. In each case it is rather the factors which should be named than the forms which are constituted by their casual collocation. In this special example of Bulimus detritus the 18 forms are made by the combinations of three pairs of independent factors. Besides these combinations which may occur anywhere or almost anywhere in the distribution there are two more distinct local forms, each of which is regarded by Coutagne as probably constituting a fresh "mode," perhaps compatible with the others.

Helix striata (Draparnauld)[13] is truly polymorphic; and its various forms have been described under various specific names. It abounds in the calcareous hills of Provence and Languedoc, disappearing in the alluvial lowlands and equally in the upper levels at about 800-1,000 m. From this district it extends through regions of similar altitude over a great part of France (details given).

Locard in his monograph of this group, which he calls collectively the group of Helix Heripensis, tabulates 27 distinct named forms. The characteristics in which these forms differ have been reckoned as 17, and as several of these vary in degree of development, the number of modes may be increased to 109. For practical purposes however Coutagne considers that the various developments of 7 characteristics in their several combinations are enough to express the various forms, and he gives examples of this method of definition. As he observes, though names may be required to define the modes, no one need be alarmed at that, for the same names of modes will be applicable to a great range of distinct species, and the formulae expressing their combinations will replace the varietal names.

This particular example of polymorphism is but little limited by locality. Occasional colonies present some special physiognomy which may in a given place seem almost invariable, though in this very respect the colonies found elsewhere may be highly variable, but such limitations are exceptional for H. striata.

Some distinct and obvious susceptibilities to the influence of soil and climate are however noticeable. For example on siliceous ground the shells are thinner, while on calcareous soils they are thicker; similarly those from the Northern districts attain a larger size than those from further South. Moreover those subjected to curtailed development, whether from drought, heat or cold often show a shortening of the spire. In contrast with this case Coutagne describes the varieties of Helix caespitum, which he says are for the most part localised, quoting many illustrative cases.

Another remarkable case in which locality plays a curious part is provided by the two species Helix trochoides and pyramidata. In France generally they are distinct enough from each other, trochoides being smaller and having a characteristic keel. Coutagne says that after having collected these species from more than a score of localities he came upon a colony of trochoides on the island of PomÈgues in which the shells were relatively enormous, most of them having only a slight keel, and a few none at all. On the other hand he received a consignment of pyramidata from four localities in Sicily, all small, and one of them exactly like the trochoides from PomÈgues. Judging by the samples received from Sicily, trochoides is there not more variable than it is in Provence, while the Sicilian pyramidata is protean.

The relations of the two species Helix nemoralis and hortensis provide an illustration of another kind of manifestation of local peculiarity. H. hortensis and nemoralis as usually met with, are two very distinct forms. H. hortensis is smaller and duller, and its peristome is white. H. nemoralis is larger and more shiny, and its peristome is brown. In several anatomical points, moreover, especially in the shape of the dart, there are great differences. For a full account of these peculiarities of the two forms and a discussion of their inter-relations the reader is referred to the elaborate work of A. Lang[14] who has studied them extensively and has also succeeded in experimentally raising hybrids between them. These hybrids were in a slight degree fertile with both the parent species, but up to the time of publication no young had been reared from hybrids inter se.

Coutagne describes the result of collections made in 62 French localities. Some had exclusively hortensis, some exclusively nemoralis, and in some the two were found in association. He gives details of five of these collections from which I take the following summary of the more essential facts, omitting much that is almost equally significant.

Locality A, near Honfleur. Both forms present, each sharply and normally distinguished, without any intermediates. They are thus found in many places. Coutagne instances MÜller's observations in Denmark, his own series from the Jura, etc.

Locality B. Vonges (CÔte d'Or), 242 hortensis taken at random, showed 128 with light peristomes (either more or less pinkish or quite white) and 114 with dark brown peristomes; together with 26 nemoralis all with the usual brown peristomes.

Of the hortensis 50 were in ground-colour opalescens and 1 roseus; and in shape 5 were umbilicatus.

Locality C, about 3 kilometres from B. There were found 35 hortensis, of which 20 had light peristomes and 15 brown; together with 7 nemoralis.

Of the hortensis none were opalescens; 18 were roseus and none has the shape of umbilicatus.

Locality D, about 1,200 metres from B. 147 hortensis, of which 4 had light peristomes and 143 had brown. No nemoralis were found.

None of the hortensis were opalescens or roseus, but 30 were umbilicatus.

In these localities intermediates of every grade existed between the well-characterised opalescens, roseus, or umbilicatus, and the other forms, but there were no intergrades between the other nemoralis and the smaller hortensis, about which there was no hesitation. In the next locality a very different state of things was found.

Locality E. Banks of the Yvette at Orsay (Seine-et-Oise). The actual numbers are not given, but we are told that 58 per cent. were hortensis, 33 per cent. nemoralis, and 9 per cent. intermediate. As at Honfleur, the hortensis had white peristomes, and the nemoralis brown. Coutagne's visits to this locality were in 1878 and 1880, and he calls attention to the fact that Pascal found similar intermediates in the same neighbourhood in 1873.

The two species, in Coutagne's view, when they occur together, can generally be sorted from each other with perfect confidence, and it is only in exceptional localities that these intermediates occur. Whether they are hybrids, or whether sometimes the species in their variations transgress their usual limitations is regarded both by Coutagne and by Lang as a question not yet answerable with certainty. Coutagne moreover lays stress on the fact that although each species may be easily known from the other in its own district, yet when shells from different districts are brought together it is sometimes impossible to sort them. He mentions an example of such casual intermixture occurring under natural conditions on an island in the Rhone, to which it may well be supposed that floods had brought immigrants from miscellaneous localities. This population contained a very large number of uncertain specimens, and as he says, it was much as if he were to mix the shells from his 62 localities, after which it would certainly be impossible to separate the two species again.[15]

Further evidence is given in the same treatise as to other examples of polymorphism, especially in the genus Anodonta, of which Locard made 251 species for France alone. Here again are cases like those already given, and many forms or "modes" are found restricted to special localities, while occasionally in the same locality dissimilar forms are found, collectively forming a colony, without intermediates.

Taken as a whole the evidence shows the following conclusions to be true. Local races, whether of animals or plants, may be distinguished by characters which we are compelled to regard as trivial, or again by features of such magnitude that if they were known to us only as the characteristics of a uniform species they would certainly be assumed without hesitation to be essential for its maintenance. Local forms may be sharply differentiated from the corresponding populations of other localities or they may be connected with them by numbers of intermediates. Not rarely also we find a fact which has always seemed to me of special significance, that the peculiarity of the local population or colony may show itself in a special liability to variation, and this variability may show itself in one of many degrees, either in the constant possession of a definite aberration, in a dimorphism, or in an extreme polymorphism.

At this stage attention should be called to two points. First, that when the details of the geographical distribution of any variable species are studied in that thorough and minute fashion which is necessary for any true knowledge of the interrelations of the several forms, the conception of a species invented by the popular expositions of Evolution under Selection is found to be rarely if ever realised in nature.

A species in this generalised sense is an aggregate of individuals, none exactly alike, but varying round a normal type, the characters of which are fixed in so far as they are adapted to environmental exigency. In nature, however, the occurrence of the varieties, and even the occurrence of the variability is sporadic. In one place a population may be perfectly uniform. In another it may be again uniform but distinct. In others the two forms may occur together, sometimes with and sometimes without intergrades. In some localities a sporadic variety may be an element of the population, persisting through long periods of time. In other localities there may be several such aberrations occurring together which are absent elsewhere.

Secondly, I would remind the reader that in the light of genetic analysis we know that intergrades, when they do occur, cannot be assumed to represent conditions through which the species must pass or has passed on its way to the extreme and definite forms.

Often, perhaps generally, they are nothing but heterozygous forms, and often also they are conditions corresponding with the presence of factors in their reduction-stages.

A broad survey of the facts shows beyond question that it is impossible to reconcile the mode of distribution of local forms with any belief that they are on the whole adaptational. Their peculiarities are occasionally the result of direct environmental influence, as we shall hereafter notice in certain cases, but none can attribute such sporadic and irregular phenomena to causes uniformly acting.

Writers on systematics, especially those of former generations often conjecture or assert that local distinctions are caused by "differences of climate, soil, food, etc.," in vague general terms. It is usually safe to assume that these remarks do not represent conclusions drawn from actual evidence, for only rarely can they be translated into more precise language. So thoroughly have the biological sciences become permeated with the belief that all distinctions are dependent upon adaptation, that the mere existence of definite distinctions is felt by many to be sufficient ground to warrant an assumption that these distinctions are directly or indirectly due to special local conditions. For example, Dr. J. A. Allen, who has done so much careful and valuable work in delimiting the local forms of the United States fauna, writes of the Ground Squirrels (Tamias)[16] as follows:—

"From the extreme susceptibility of this plastic group to the influences of environment, it is one of the most instructive and fascinating groups among North American mammals. No one can doubt its comparatively recent differentiation from a common stock, and its dispersion from some common centre. Whether the type originated at some point in North America, or in the Northern part of Eurasia, it is perhaps idle to speculate, but that it has increased, multiplied, spread, and become differentiated to a wonderful degree in North America is beyond question; as it is found from the Arctic regions to the high mountain ranges of Central Mexico, and has developed some twenty to thirty very palpable local phases."

"Some of them easily take rank as species, others as subspecies. Probably a more striking illustration of evolution by environment cannot be cited."

He proceeds to point out that the habits of these creatures are such as lead to isolation. This may well be admitted, and indeed no exception can possibly be taken to the passage as a whole, save in the one respect that there is no real proof that the local diversity is due to "evolution by environment" or an indication of "susceptibility to the influences of environment."

Dr. Allen does indeed adduce the fact that California "extending through 800 miles of latitude, with numerous sharply contrasted physiographic regions, has apparently no less than six strongly differentiated forms, while the region east of the Rocky Mountains from a little below the northern boundary of the United States northward to the limit of trees—a slightly diversified region of at least ten times the area of California—has only one"! But when one comes to ask how the various forms are adaptational, and how the influences of environment have led to their production, only conjectures of a preliminary and tentative character could be expected in reply. Desert forms are no doubt pallid as in so many instances, and forest forms are more fully coloured, and we may readily enough accept such facts as indications of a connection between bodily features and the conditions of life, but further than that no one can go; so that when we find size, length of ears or of tail, the number of dorsal stripes, the pattern of the colours, not to speak of differences in the pigments themselves, all exhibiting large modifications, we cannot refer these peculiarities to the causation of environmental difference, save as a simple expression of faith. I incline far more to agree with Gulick who, after years of study of the local variations of the Achatinellidae, came to the conclusion that it was useless to expect that such local differentiation can be referred to adaptation in any sense.[17] Even the most convinced Selectionist must hesitate before such facts as those related by A. G. Mayer regarding the distribution of Partula otaheitana, one of these Achatinellidae. The island of Tahiti has been scored by erosion so that a series of separated valleys radiate to the coast. From four successive valleys Mayer collected the species, and found that in the first (Tipaerui) valley all the shells were dextral (115, containing 73 young); in the second valley (Fautaua) 54 per cent. of adults and 55.5 per cent. of the young contained were sinistral; in the third valley (Hamuta) 69 per cent. of adults and 73 per cent. of young contained in them were sinistral; and lastly, in the fourth valley (Pirae) all the shells (131, containing 62 young) were sinistral.[18] In connection with these observations I may mention the fact that in a certain pond in the North of England[19] the sinistral form of Limnaea peregra has been known to occur for about fifty years. Visiting it lately I found the left-handed shells to be about 3 per cent. of the population. The species is the commonest British freshwater shell, but left-handed specimens are exceedingly rare. Will anyone ask us to suppose that the persistence of a percentage of this rarity in the same place is an indication of some specially favouring circumstance in the waters of that pond? It is a horse-pond to all appearances exactly like any other horse-pond; and I believe that in perfect confidence we may accept the suggestion of common sense, which teaches us that there is nothing particular in the circumstances which either calls such varieties into existence or contributes in any direct way to their survival. Had the phenomenon of local variation been studied in detail before Darwin wrote, the attempt to make selection responsible for fixity wherever found, could never have been made. The proposition that not only the definiteness of local forms but their variability also is sporadic, can be established by countless illustrations taken from any group of either the animal or the vegetable kingdoms. Only exceptionally can the fixed differences be even suspected of contributing to adaptation, and sporadic variability, which is a no less positive fact, must manifestly lie outside the range of such suspicions. It is open to any one to suggest speculatively that the persistence of special varieties or of special variability in special places is an indication that in those places the conditions of life are such that the forms in question are tolerated though elsewhere the same types are exterminated; but that consideration, even if it could be proved to be well founded, is not one which lends much force to the thesis that definiteness of type is a consequence of Natural Selection. On the contrary, recourse to such reasoning implies the inevitable but very damaging admission that the stringency of Selection is frequently so far relaxed that two or more equally definite forms of the same species can persist side by side. There is no doubt that this is the simple truth, but when once that truth is perceived it is useless to invoke the control of Selection as the factor to which definiteness of type in general must be referred.

The genetic relations of local forms to each other cannot in the absence of actual breeding experiments be often ascertained. Standfuss formerly enunciated as a general principle that when two forms co-exist in the same locality and are able to interbreed, they do not produce intermediates; but that when the forms are geographically separated as local races, crosses between them result in a series of intermediates.[20] In this aphorism there is a good deal of truth, but if in the light of Mendelian principles we examine the two statements we see now that the first is in reality only another way of saying that the distinctness of an aberrational form co-existing with another is due to segregation, accompanied by some degree of dominance of one type. Whether, however, one geographically isolated race will give intermediates when bred with another must depend entirely on the genetic physiology of the special case, and no general rule can be laid down. It may well be that, inasmuch as the distinctness of the variety is maintained by isolation, the difference in factorial composition between it and the representative form in another area is neither simple nor sharp; but when two varieties co-exist, though interbreeding, it is now clear that their differences must depend on the segregation of simple factors. Plainly such aberrations may in one place co-exist with another type, and elsewhere be separated from it as local races.

Excellent illustrations of these two stages in evolution are provided by the melanic varieties of British Lepidoptera. The fact that black or blackish varieties of many species especially of Geometridae have come into existence in recent years is well known to British collectors, and it is not in dispute that they have in several instances replaced the older type more or less completely in certain districts. In the year 1900 the Evolution Committee of the Royal Society instituted a collective inquiry as to the contemporary distribution of these dark varieties. As the change had happened within living memory and had greatly progressed in recent years it was hoped that a record of the existing distribution would serve as a point of departure for future comparison. The records thus obtained were tabulated by Mr. L. Doncaster.[21] From that account and from the statements in Barrett's British Lepidoptera[22] this description of some of the more notable cases is taken.

The most striking and familiar case is that of Amphidasys betularia, of which only the ordinary type was known in any locality until about 1848-1850, when the totally black var. doubledayaria first appeared in the neighbourhood of Manchester. This black form was subsequently recorded in Huddersfield between 1860 and 1870; Kendal about 1870; Cannock Chase, 1878; Berkshire, 1885; Norfolk, Essex and Cambridge about 1892; Suffolk, 1894; London, 1897. For the Southern Counties of England, except in the London district, there are still very few records. It cannot of course be asserted positively that the variety spread from its place of first appearance into the other localities, and that it did not arise de novo in them, but there can be little doubt that the process was one of colonisation. On the European Continent the first records are from Hanover in 1884, Belgium 1886 and 1894, Crefeld 188-, Berlin 1903, Dresden about the same date.

As regards the increase of the variety we have the fact that in Lancashire, Cheshire and the West Riding of Yorkshire the black is now the prevalent form; and in some places, as for example, Huddersfield, the black alone is now found, though it was unknown there till between 1860 and 1870. About 1870 at Newport, Monmouth, the two forms were in about equal numbers, but a few years later the type had almost vanished. Similarly in Crefeld, where the black form was still very rare in the eighties, it now forms about 50 per cent. of the population. In the London district the black remains scarce and at the date of the report it was still very scarce. From Ireland there is only one record and there are hardly any from Scotland.

Boarmia repandata is another species which is behaving in a somewhat similar way. Unlike betularia, however, the species is a variable one, and has several colour-forms, amongst them the banded var. conversaria, and many others. In addition to these there is a black form in the North of England which seems to be spreading. In Huddersfield the black was first recorded in 1888, and in 1900 20-25 per cent. were black. At Rotherham the black or very dark are now prevalent and have increased in the last 15 years. From the Midlands, East Anglia and Southern Counties the returns show only the light and medium forms.

Of Odontoptera bidentata several intergrading dark forms exist, and these are found exclusively in the North and the Midlands. Unicolorous blacks have been found recently in the Lancashire mosses and at Wakefield. At Huddersfield 50 years ago the light forms were prevalent, but now a rather dark brown, not infrequently suffused with black, is the commonest. In Southern Counties only light forms are known.

Phigalia pilosaria in South England is always light, but in the North the prevalent form is darker. About 35 years ago a form with unicolorous sooty fore-wings and dull grey hind wings was first seen in Yorkshire and a similar form is now taken regularly in South Wales.

In the following cases the dark varieties were found originally only in the South.

Boarmia rhomboidaria gave rise about 40 years ago to a unicolorous smoky variety called perfumaria. This was at first peculiar to the London district, but it has since been taken in Birmingham and other large cities. More lately coal-black specimens have been found at Norwich, and others similar but hardly so dark were taken in the South of Scotland and at Cannock Chase.

Eupithecia rectangulata is a similar case. Formerly the light forms were prevalent but within sixty years they have almost entirely been replaced in the South of London by a nearly black form.

Tephrosia (Boarmia) consortaria and Tephrosia consonaria are exceptionally interesting, for they have both given off dark forms in the same wood near Maidstone, which is far from the usual "centres of melanism." They were discovered in this locality by Mr. E. Goodwin. That of consortaria is a dark grey, but that of consonaria is a full black, and nothing like either has been found anywhere else.

These examples are all taken from the Geometridae but others, though of a less conspicuous kind, could be given from the Noctuidae or the Micro-Lepidoptera. Acronycta psi, for instance, has a suffused form which is believed to be becoming more frequent in the London district. Polia chi has two dark forms, olivacea, a yellowish grey with dark markings, and suffusa which is a darker, blackish-slate colour. Both occur in the North of England, sometimes together, sometimes separately, or mixed with the type and many intermediates. The distribution is peculiarly irregular. At Huddersfield, where the very dark form appeared suddenly about 1890, some 30 per cent. are said to be now dark and about 6-7 per cent. very dark, but at Saddleworth, 12 miles away, only the pale forms occur.

Several questions of interest arise in regard to this evidence. This progressive Melanism has arisen in certain families only, and may be confined to certain species only, within those families. As in almost all other examples in which variation has been much observed, its incidence is capricious and specific. A collateral line of inquiry relates to the degree of discontinuity which the variation manifests. Here again there is no rule. Generally speaking, in A. betularia, to take the case most fully studied, the variation is discontinuous. Real intermediates between betularia and doubledayaria are in most localities absent or rare. The black spots of betularia may often be larger or more numerous than in the normal, but this variation has nothing to do with doubledayaria, and is not an intermediate stage towards it, though sometimes wrongly so described. Doubledayaria owes its characteristic appearance to a factor which blurs the surface of the wings with a layer of black. Sometimes this blurring is slighter than in the real doubledayaria, and these forms are real intermediates. Occasionally the fore-wings alone are thus blurred. These intermediates are clearly due to reduction-stages of the doubledayaria factor, and are related to it as a blue mouse is to a black, or a dutch rabbit to a self-colour. It cannot positively be asserted that the full doubledayaria existed before the intermediate, but it almost certainly did. In certain places as for instance in Belgium, there is evidence that intermediates have at various times been fairly abundant, but they have never become common, nor are they known to exist in the absence of doubledayaria. When the black variety and the light type breed together they do not usually have intermediates among their offspring, and the evidence is consistent with the view that the black is a complete dominant. The same is probably true of Tephrosia consonaria.

In some of the other species we know that the darkest forms did not appear first. For example in Phigalia pilosaria and Boarmia rhomboidaria dark forms existed and are believed to have increased in number before the darkest made its appearance. Hybernia progemmaria is said to have become darker gradually both in Cheshire and in the West Riding, and a uniformly smoky variety appeared in South Yorkshire less than 45 years ago which has spread to neighbouring counties. The dark medium has become the commonest form in Huddersfield district, where the very dark variety is now about 20 per cent. of the population, though the light form is still common.

Taking the evidence together we find it consistent with the view that dark forms have appeared sporadically, in some species the very dark appearing first and intermediates later, in others the moderately dark came first and the darkest later in time. It is practically certain that the change has in general come about not by a gradual change supervening on the population at large, but by the sporadic appearance of dark specimens as a new element in the population, and strains derived from these dark individuals have gradually superseded the normal type more or less completely.

If it could be shown that these melanic novelties had a definite advantage in the struggle for existence they would provide an instance of evolution proceeding much in the way which Darwin contemplated. The whole process would differ from that conceived by him as the normal method of evolution only in so far as the change has come about with great rapidity and in some instances largely by the appearance and success of discontinuous varieties. The question, however, must be asked whether the dark form can reasonably be supposed to have an advantage by reason of their darkness. Some naturalists believe that the darkness of the colours does thus definitely contribute to their protection by making the insects less conspicuous and thus more likely to escape the search of birds. In support of this view it may be pointed out that it is in the manufacturing districts of Lancashire and Yorkshire, and again in the London area that the melanics have attained their greatest development. Consistently with this argument also, it is in the neighbourhood of Crefeld and Essen, the black country of Germany, that they have chiefly established themselves on the Continent, and Phigalia pilosaria in the black form is now at home in South Wales. Thus superficially regarded, the evidence looks rather strong, but it is difficult to apply the reasoning in detail. We have first the difficulty that the black form of betularia for instance has established itself in thoroughly rural districts, notably near King's Lynn in Norfolk, and in the neighbourhood of Kendal and Windermere. The black form of consonaria and the dark consortaria appeared in a wood near Maidstone, far from town smoke, and the black rhomboidaria was first found at Norwich, which, as towns go, is clean. Then again the spread of the melanics is very irregular and unaccountable. The black pilosaria is found both in the West Riding and in the Swansea district, but not yet elsewhere. It rapidly increased at Huddersfield, but made no noticeable progress at Sheffield though recorded there for ten years. It is also a remarkable fact that no similar melanic development has been observed in America, and, so far as I am aware, comparable melanic varieties have not appeared on the European continent except in the case of the few sorts which possibly may have come from England.

The whole subject is beset with complications. It must not be forgotten that in a few species of moths there is an obvious and recognised conformity between the colours of the perfect insect and that of the soil on which they live, comparable with that which is so striking in the case of some Oedipodidae and other grasshoppers. Of this phenomenon the clearest example is Gnophos obscurata, which is a most variable species with many local forms. Of these a well-known dark variety lives on the peaty heaths of the New Forest and other districts, but on the chalk hills of Kent, Sussex and Surrey various light varieties are found, of which one is a bright silvery white, very near in colour to the colour of a chalky bank. This case does not seem to be one of direct environmental action,[23] for Poulton found no change induced by rearing larvae among either white or black surrounding objects. No one however can doubt that there is some indirect connection between the colour of the ground and that of the moths.

To my mind there is a serious objection to the theory of protective resemblance in application to such a case as that of the betularia forms, which arises from the fact that the black doubledayaria is a fairly conspicuous insect anywhere except perhaps on actually black materials, which are not common in any locality. Tree trunks and walls are dirty in smoky districts but they are not often black, and I doubt whether in the neighbourhood of Rotherham, for instance, which is one of the great melanic centres, doubledayaria can be harder for a bird to find than betularia would be. After all, too, many of the species much affected are not urban insects. They live in country places between the towns, and the general tone of these places even in Lancashire and the West Riding is not very different from that of similar places elsewhere. As against the objection that the black varieties are much blacker than the case requires it may be replied that we know nothing of the senses of birds, and that perhaps to their eyes blackness does constitute a disguise even though the surroundings are much less dark. This is undeniable, but recourse to such an argument is dangerous; for if the sight of the insect-eating birds is so dull that it does not distinguish dark things from dingy grey, we cannot subsequently regard the keen sight of birds as the sufficient control which has led to the minute and detailed resemblance of many insects to their surroundings. Those who see in such cases examples of the omnipotence of Selection must frequently find themselves in this dilemma.

Taking the evidence as a whole, we may say that it fairly suggests the existence of some connection between modern urban developments and the appearance and rise of the melanic varieties. More than that we cannot yet affirm. It is a subject in which problems open up on every side, and all of them are profitable subjects for investigation. Unhappily such animals are difficult to rear successfully in captivity for many generations, owing to their extreme liability to disease. Not the least interesting feature of the melanics is the fact that the black varieties provide about the best and clearest example of a new dominant factor attaching itself to a wild species in recent times. None of the cases are satisfactorily recorded or analysed as yet, but the evidence is clear that doubledayaria is a dominant to its type, and in several other dark varieties, though the pigment deposited is not black, the records show that the increased amount of the pigment almost certainly is due to a positive factor. Of this, Hemerophila abruptaria is a good example.[24] There are some irregularities in the results, but taken together they leave little doubt that the dark brown variety is a dominant and the light, yellowish brown a recessive.

A curious parallel to the rise of the melanic moths in England is provided by the case of the Honey-creepers or Sugar-birds, in certain West Indian islands.[25] These birds of the genus Coereba (Certhiola) range from Southern Mexico to the Northern parts of South America and through the whole chain of the West Indian islands and Bahamas except Cuba. There are numerous local forms, and many of the islands have types peculiar to themselves, as is usual in such cases. Some of the types or species range through several islands, but according to Austin Clark[26] no island has more than one of them. Cory[27] reckoned twelve such species within the Antillean region. They are small birds about the size of a nuthatch with a general colouring of black, yellow, and white. From the island of St. Vincent the Smithsonian Institution received in the late seventies of last century several completely black specimens in addition to two of the usual type of colouring. The black were described by W. N. Lawrence as atrata, and those marked with the usual yellow and white were called saccharina. The collector (Mr. F. A. Ober) reported that the black form was common, and that the saccharina form was rarer. Lawrence remarks, "Had there been only a single example (of the black form) I should have considered it as probably a case of abnormal colouring, but it seems to be a representative form of the genus in this island."[28] There is of course no doubt of the correctness of the view taken by Austin Clark that "atrata" is a black variety. The black bird is in every respect, other than colour, identical with saccharina, and it is even possible to detect a greenish colour in the areas which would normally be yellow, showing plainly enough the yellow pigment obscured by the black.

We have next the interesting fact that like our melanic moths the dark form is replacing the "type." At the time of Ober's visit the type was already in a minority, but now it is nearly or perhaps actually extinct, though the black form is one of the commonest birds on the island. Austin Clark found no specimen when he collected there in 1903-4, though formerly it was not uncommon in the vicinity of Kingston and in the immediate windward district of St. Vincent.

The Grenadines are geographically just south of St. Vincent, though separated by a deep channel. In these islands no black forms have yet been taken, but Grenada, the next island to the south, has both normals and blacks. There are trifling differences of size between the Grenada birds and those from St. Vincent, the Grenada specimens being slightly smaller and for this reason they have received distinct names, the form marked with yellow and white being called Godmani (Cory) and the black, Wellsi (Cory), but this merely introduces a useless complication. There is evidence that in Grenada, as in St. Vincent, the black is gradually ousting the original type, but the process has not gone so far as in St. Vincent. Austin Clark very properly compares this case of the Sugar-birds with that of Papilio turnus, which as is well-known, has a black female in the southern parts of its distribution, in addition to a female of the yellow type, but in the Northern States the black female does not occur.

During the present year P. R. Lowe, who lately studied Coerebas on a large scale in the West Indies, has published an important paper on the subject.[29] He calls attention to the fact that Cory recently found a black form of Coereba on Los Roques Islands, and he himself discovered another on the Testigos Islands. Both localities are on the coast of Venezuela, far from St. Vincent and Grenada. The whole problem is thus further complicated by the fact that the black varieties have, as we are almost driven to admit, arisen independently in remote places. Improbable as this conclusion may be, it is still more difficult to regard all the black forms as derived from one source. For first, they present definite small differences from each other; and secondly we have to remember a consideration of greater importance, that the very fact that each island has its own type must be accepted as proving that the localities are effectively isolated from each other, and that migration must be a very rare event.

The rarity of such illustrative cases is, I believe, more apparent than real. It is probably due to the extreme reluctance of systematists to admit that such things can be, and of course to the almost complete absence of knowledge as to the genetic behaviour of wild animals and plants. Only in such examples as this of the Coereba, where colour constitutes the sole difference, or that of the moths which have been minutely studied by many collectors, does the significance of the facts appear. The arrangement of catalogues and collections is such that much practical difficulty of a quite unnecessary kind is introduced. For example, in this very case of Coereba, I find the British Museum has a fine series from Grenada including 3 normals and 11 black, and also 16 blacks from St. Vincent. If the black specimens from Grenada were put with the normals which are almost certainly nothing but a recessive form of the same bird, the variation would strike the eye on even a superficial glance at the drawer. But following the notions so naively expressed in the passage quoted above from W. N. Lawrence, the blacks from Grenada are put apart together with the other blacks from St. Vincent, though two of them were shot on the same date as one of the normals.