CHAPTER VII

LOCAL DIFFERENTIATION. _Continued_

OVERLAPPING FORMS

The facts of the distribution of local forms on the whole are consistent with the view that these forms come into existence by the sporadic appearance of varieties in a population, rather than by transformation of the population as a whole. Of such sporadically occurring varieties there are examples in great abundance, though by the nature of the case it can be but rarely that we are able to produce evidence of a previous type being actually superseded by the variety. When the two forms are found co-existing in the same area they are usually recorded as one species if intergrades are observed, and as two species if the intergrades are absent. On the other hand when two forms are found occupying separate areas, when, that is, the process of replacement is completed in one of the areas, then forthwith each is named separately either as species or subspecies. Successive observations carried out through considerable periods of time would be necessary to establish beyond question that the history proceeds in one way rather than another. Such continuity of observation has for the most part never been attempted. The kind of information wanted has indeed only been lately recognized, and really critical collecting is a thing of only the last few decades. The methods of the older collectors, who aimed at bringing together a few typical specimens of all distinct forms, are of little service in this class of inquiry, which is better promoted by the indiscriminate collection of large numbers of common forms from many localities. When this has been done on a comprehensive scale we shall be in a position to form much more confident judgments as to the general theory of evolution.

Some little work of the kind has however been done and the results are already of great value. Seeing that the differentiation of local forms is only made possible by isolation, it necessarily happens that the collector finds one form in one locality and another in a distinct locality, and there is no evidence as to the behaviour which the two representative species might exhibit if they came into touch with each other. In the most familiar examples of such distinction each inhabits an island, completely occupying it to the exclusion of any other similar form. It can only be when the two representative species occupy parts of a continental area connected with each other by regions habitable for the organism in question, that there is a chance of seeing the two forms in contact. Often also, even where this condition is satisfied, the habits, social organisation, or some other special cause may act as a barrier which prevents the distinguishable forms from ever coming into such complete contact as to interbreed or to behave as a genetically continuous race. When genetic continuity is ensured by a constant diffusion of the population over the whole area which they inhabit there will manifestly be no formation of local races. The practical uniformity, for example, of so many species of birds which inhabit widely extended ranges of Western Europe is doubtless maintained by such constant diffusion. When, as in the case of the Falcons, many localities have peculiar forms, the fact may be taken as conclusive evidence that there is little or no diffusion; and when we find in such a species as the Goldfinch that in spite of migratory fluctuations there are nevertheless geographical races fairly well differentiated, it may similarly be inferred that these fluctuations habitually move up and down on paths which do not intermingle. There are however a few examples of animals, not given to much irregular wandering, which occupy a wide and continuous range of diversified country and are differentiated as local races in two or more districts, though the distinct races meet in intervening areas. Of these the most notorious illustration which has been investigated with any thoroughness is that of the species of _Colaptes_ (Woodpeckers) known in the United States as Flickers. The study of the variations of these forms, made by J. A. Allen[1] is an admirable piece of work, with which every student of variation and evolutionary problems should make himself familiar. The two forms with which we are most concerned are known as _C. auratus_ and _C. cafer_, and are very strikingly different in appearance. In size, proportions, general pattern of colouration, habits, and notes, the two are alike, but they differ in the following seven respects as stated by Allen.

_Auratus_ _Cafer_

1. Quills _yellow_. 1. Quills _red_. 2. Male with a _black_ malar 2. Male with a _red_ stripe. malar stripe. 3. Adult female with _no_ 3. Adult female with usually a malar stripe. brown malar stripe. 4. _A scarlet nuchal crescent 4. No nuchal crescent in in both sexes._ either sex. 5. Throat and fore neck 5. Throat and fore neck _brown_. _grey_. 6. Whole top of head and hind neck 6. Whole top of neck and hind _grey_. neck _brown_. 7. General plumage with an 7. General plumage with a _olivaceous_ cast. _rufescent_ cast.

These differences are illustrated in the accompanying coloured plate, which has been most kindly prepared for me under the instructions of Dr. F. M. Chapman of the American Museum of Natural History. Before going further it is worth considering the nature of these differences a little more closely. All but the last are large differences which no one would overlook even in a hasty glance at the birds. If the only distinction lay in the colour of the quills we might feel fairly sure that _auratus_ was a recessive form of _cafer_, and so probably it is in this respect. Similarly the black malar stripe of _auratus_ is in all probability recessive to the red malar stripe of _cafer_ and I imagine the pigments concerned are comparable with those in the Gouldian Finch (_Poephila gouldiae_) of Australia. Both sexes in that species may have the head black, red, or, less often, yellow, and though it is not any longer in question that birds may breed in either plumage, I believe that the young are always black-headed and I imagine that those which become red-headed possess a dominant factor absent from the permanently black-headed birds.[2] Yellow as a recessive form of a red is certainly very common, but red and black as variants of the same pigment are less usual. In the Gouldian Finch we seem to have a case where a pigment can assume all three forms. It would be interesting to know whether the red of the malar stripes in _Colaptes_ is a pigment of the same nature as the red of the quills. Both in _Colaptes_ and in _Poephila gouldiae_ I have seen specimens intermediate between the black and the red, and the appearance of the part affected was exactly alike in the two cases, red feathers coming up among the black ones, and many feathers containing both red and black pigments mixed together. The development of the scarlet nuchal crescent in _auratus_ and the absence of this conspicuous mark in _cafer_ constitute from the physiological point of view the most remarkable pair of differences. When the red crescent is not formed, the feathers which would bear it are exactly like the rest, and no special pigment is visible in them which one can regard as ready to be modified into red. If the crescent is due to a factor it must therefore be supposed that this factor has the power of modifying the pigment of the neck in one special place alone. Dr. W. D. Miller called my attention to the fact that a similar variation occurs in another American woodpecker, the Sapsucker, _Sphyropicus varius_.[3]

I do not suggest that such variations are without parallel: indeed in _P. gouldiae_ the factor which turns the black of the head into scarlet affects one special region of the black only, being sharply distinct from the unmodified black of the throat. These regions of the head are however often the seat of special colours in birds.[4] So also may be instanced the variety of the Common Guillemot (_Uria troile_) which has a white line round the eyes and at the sides of the head where the normal has no such mark; but this line is formed in a very special place, the groove joining the eye to the ear, whereas the feathers of the nuchal crescent are not ostensibly distinguished from those adjacent.[5]

The transposition of the brown and the grey on the back and front of the neck also constitutes a very remarkable difference. If either grey or brown depends on a factor then it must be supposed that _auratus_ has one of these factors and _cafer_ the other.

From these several considerations it is quite clear that if _auratus_ and _cafer_ are modifications of the same type produced by presence or absence of factors, several independent elements must be concerned, and to unravel their inter-relations would be most difficult even if it were possible to breed the types under observation, which is of course quite beyond present possibilities.

The distribution of the two is as follows. On the east side of the Continent _C. auratus_, relatively pure, occupies the whole of Canada and the States from the North to Galveston. Westward it extends across the whole continent in the more northern region to Alaska, but in its pure form it only reaches down the Pacific coast to about the northern border of British Columbia. Its southern and western limit is thus roughly a line drawn from north of Vancouver, southeast to North Dakota and then south to Galveston. _C. cafer_ in the comparatively pure form inhabits Mexico, Arizona, California (except Lower California and the opposite coast), central and western Nevada, Utah, Oregon, and is bounded on the east by a line drawn from the Pacific south of Washington, south and eastward through Colorado to the mouth of the Rio Grande or the Gulf of Mexico. Between the two lines thus roughly defined is a band of country about 1,200-1,300 miles long and 300-400 miles wide, which contains some normal birds of each type, but chiefly birds exhibiting the characters of both, mixed together in various and irregular ways. Even in the areas occupied by the pure forms occasional birds are recorded with more or less indication of characteristics of the other form, but within the area in which the two forms are conterminous, the mixed birds are in the majority. The condition of these birds of mixed character is described by Allen as follows:

"As has been long known--indeed, as shown by Baird in 1858--the 'intermediates' or 'hybrids' present ever-varying combinations of the characters of the two birds, from individuals of _C. auratus_ presenting only the slightest traces of the characters of _C. cafer_, or, conversely--individuals of _C. cafer_ presenting only the slightest traces of the characters of _C. auratus_--to birds in which the characters of the two are about equally blended. Thus we may have _C. auratus_ with merely a few red feathers in the black malar stripe, or with the quills merely slightly flushed with orange, or _C. cafer_ with either merely a few black feathers in the red malar stripe, or a few red feathers at the sides of the nape, or an incipient, barely traceable scarlet nuchal crescent. Where the blending of the characters is more strongly marked, the quills may be orange-yellow or orange-red, or of any shade between yellow and red, with the other features of the two birds about equally blended. But such examples are exceptional, an unsymmetrical blending being the rule, the two sides of the same bird being often unlike. The quills of the tail, for example, may be part red and part yellow, the number of yellow or red feathers varying in different individuals, and very often in the opposite sides of the tail in the same bird. The same irregularity occurs also, but apparently less frequently, in the quills of the wings. In such cases the quills may be mostly yellow with a few red or orange quills intermixed, or red with a similar mixture of yellow. A bird may have the general colouration of true _cafer_ combined with a well-developed nuchal crescent, or nearly pure _auratus_ with the red malar stripes of a _cafer_. Sometimes the body plumage is that of _C. auratus_ with the head nearly as in pure _cafer_, or exactly the reverse may occur. Or we may have the general plumage as in _cafer_ with the throat and crown as in _auratus_, and the malar stripe either red or black, or mixed red and black, and so on in almost endless variations, it being rare to find, even in birds of the same nest, two individuals alike in all their features of colouration. Usually the first trace of _cafer_ seen in _auratus_ manifests itself as a mixture of red in the black malar stripe, either as a few red feathers, or as a tipping of the black feathers with red, or with merely the basal portion of the feathers red. Sometimes, however, there is a mixture of orange or reddish quills, while the malar stripe remains normal. In _C. cafer_ the traces of _auratus_ are usually shown by a tendency to an incipient nuchal crescent, represented often by merely a few red-tipped feathers on the sides of the nape; at other times by a slight mixture of black in the red malar stripe."

Such a state of things accords very imperfectly with expectations under any received theory of Evolution. As in some of the instances discussed in the first chapter we have here two fairly definite forms, nearly allied, which on any evolutionary hypothesis must have been evolved either the one from the other, or both from a third form at a time not very remote from the present, as time must be measured in evolution. Yet though intermediates exist in some quantity, no one can for a moment suggest that they are that definite intermediate from which _auratus_ and _cafer_ descend in common. One cannot imagine that the immediate ancestor of these birds was a mosaic, made up of asymmetrical patches of each sort: but that is what many of the intermediates are. It is not much easier to suppose the ancestor to have been a nondescript, with a compromise between the developed characters of each, with quills buff, malar stripes neither black nor red, with a trace of nuchal crescent, and so on. Such Frankenstein-monsters have played, a considerable part in the imaginations of evolutionary philosophers, but if it were true that there was once a population of these monsters capable of successful existence, surely they should now be found as a population occupying the neutral zone between the two modern forms. Yet, though much remains to be done in clearing up the facts, one thing is certain, namely that the neutral zone has not a definite and normally intermediate population, but on the contrary it is peopled by fragments of the two definite types and miscellaneous mongrels between them.

On the other hand, one cannot readily suppose that either form was the parent of the other. The process must have involved both addition and loss of factors, for whatever hypothesis be adopted, such changes must be supposed to have occurred. A careful statistical tabulation of the way in which the characters are distributed in the population of the mixed zone would be of great value, and till that has been done there is little that can be said with certainty as to the genetics of these characters. In the collection of Dr. Bishop of New Haven I was very kindly allowed to examine a sample, all taken at random, near together, in Saskatchewan. There were females 4 adult, 2 young; males 4 adult and 5 young. This number, though of course insufficient, is enough to give some guide as to the degree of definiteness which the characters generally show in their variations. Of the 15 birds, 8 had simply yellow quills; 2 had red; 1 was almost red but had one yellow tail-quill; 3 were intermediate and 1 was buff. As regards the malar patch, which can only be determined properly in the adult males, 1 was red, 1 was approximately red, 2 intermediate. As to nuchal crescent 4 females had none, 2 females very slight; 7 males had it, 1 had only a slight crescent, and 1 had none. In point of quills therefore 10 were definite out of 15; in point of crescent, 11 were definite out of 15; and in point of malar patch 1 only was definite out of 4. The last is a feature directly dependent on age and so counts for less, but as regards the other two features there is some indication that the factors show definiteness in their behaviour. It must be remembered that we have no knowledge what the heterozygous form may be, and in the case of red and yellow it is probably a reddish buff. The patch-works are no doubt to be compared with other well-known pied forms, and in these we must suppose the active factor broken up, which it probably can be very easily. The asymmetry, which Allen notices as so marked a feature, in the distribution of the red and yellow quills of the tail especially, recalls that of the black markings in the pied Canaries. As is well known to students of variations _some_ pigment-factors in _some_ animals are apparently uncontrolled by symmetry, while in other specific cases symmetry is the rule. On the other hand the blackness or redness of the malar patches is, I think, as a rule nearly symmetrical. It should be mentioned that two of Dr. Bishop's young birds belonged to the same nest, one a female with _red_ quills, the other a male with _yellow_. Both are without crescent.

As to the question whether certain combinations of characters occur with special frequency, the evidence is insufficient to give a definite answer. Among all the birds I have seen in America or in England I have not yet found one having the malar patches black without any nuchal crescent. Of Dr. Bishop's 8 adults not one, however, showed the combination of the three chief features normal for _auratus_ or for _cafer_.

Besides the two forms that we have hitherto considered, several other local types exist, and these throw some further light on the problem. Of these the most important in this connexion is _chrysoides_, which inhabits the whole of southern California and the mainland opposite. This remarkable form is as Allen says, very different from _auratus_ except that it has the quills yellow like _auratus_, not red like _cafer_. So that we find here in the extreme west of the whole distribution a type agreeing in one of its chief features with the eastern type. Between this and _cafer_ intergrades have, according to Allen, not been found. The relations of this _chrysoides_ are, Allen thinks, rather with _mexicanoides_, a southern, smaller race with colours more intense, which inhabits Guatemala, but however that may be, it must be regarded as a _cafer_ which has lost its red quills. The island of Guadeloupe off Lower California has an island form. Beyond the other side of the continent there is also an island form of _auratus_, inhabiting Cuba, so that clearly the yellow quills can extend into the tropics.

The above account is in many respects incomplete, but it suffices to give an outline of the chief facts. The whole problem is complicated by the undoubted effects of an uncertain amount of migration, and in many, perhaps all, districts, the winter population differs from the summer population of the same localities. The existence of these seasonal ebbs and flows is now well known to ornithologists, and most of the bird species of temperate regions are subject to them.

Difficult as it may be to conceive the actual process of origin of the two types _auratus_ and _cafer_, it is I think still harder to suggest any possible circumstance which can have determined their development as distinct races, or which can maintain that distinctness when created. Some will no doubt be disposed to appeal once more to our ignorance and suggest that if we only knew more we should see that the yellow quills, the black "moustache" and the red crescent, specially qualify _auratus_ for the north and eastern region, and the red quills, red "moustache" and absence of crescent fit _cafer_ to the conditions of its homes. Each can judge for himself, but my own view is that this is a vain delusion, and that to cherish it merely blunts the receptivity of the mind, which if unoccupied with such fancies would be more ready to perceive the truth when at last it shall appear. Think of the range of conditions prevailing in the country occupied by _auratus_--a triangle with its apex in Florida and its base the whole Arctic region of North America. Is it seriously suggested that there is some element common to the "conditions" of such an area which demands a nuchal crescent in the Flickers, though the birds of the _cafer_ area, almost equally varied, can dispense with the same character? Curiously enough, the geographical variation of _Sphyropicus varius_, another though a very different Woodpecker[6] shows that conversely the nuchal crescent can be dispensed with in the Eastern form though it is assumed by the Western.[7]

Allen points out the interesting additional fact that superposed upon each of the two distinct forms, _auratus_ and _cafer_, are many geographical variations which can very naturally be regarded as climatic. Each decreases in size from the North southward, as so many species do.[8] They become paler in the arid plains, and show the ordinary phases which are seen in other birds having the same distribution. Such differences we may well suppose to be determined directly or indirectly, by environment, and we may anticipate with fuller knowledge it will be possible to distinguish variations of this nature as in the broad sense environmental, from the larger differences separating the two main types of _Colaptes_, which I surmise are altogether independent of such influences.

It is generally supposed that phenomena like those now so well established in the case of _Colaptes_ are very exceptional, and as has already been stated a number of circumstances must combine in order that they may be produced. I suspect however that the examples are more numerous than is commonly thought. In all likelihood the three forms _Sphyropicus varius_, _nuchalis_ and _ruber_ are in a very similar condition though the details have not, so far as I know, been worked out. A complex example which is closely parallel to the case of _Colaptes_ was described by F. M. Chapman[9] at the same date as Allen's work. This is the case of _Quiscalus_, the Grackles, which in the North American Continent have three fairly distinct forms which Chapman speaks of as _Q. aeneus_, _Q. quiscula_, and _Q. quiscula aglaeus_. The birds are all, so far as pigment is concerned, dark blackish brown, but the head and mantle have superposed a metallic sheen of interference-colours which in the various forms take different tints, bluish green, bronze green, or bronze purple. The details are complicated and difficult to appreciate without actual specimens, but the two common types are sufficiently distinct. The birds inhabit the whole area east of the Rockies, _quiscula aglaeus_ occupying Florida and the Southern States southwest of a band of country about a hundred miles broad extending roughly from Connecticut to the mouth of the Mississippi; and _aeneus_ taking the area north and west of this band. In discussing this case Chapman expresses the same view as Allen does in the _Colaptes_ case, that there are two distinct populations, substantially fixed, and that the band of country in which they meet each other has a mongrel population, with no consistent type, but showing miscellaneous combinations of the character of the two chief types.

The warblers of the genus _Helminthophila_ provide another illustration which has points of special interest. The two chief species are _H. pinus_, which has a yellow mantle and lower parts, white bars on the wings, a black patch behind the eyes and a broad black mark on the throat; and _H. chrysoptera_ with dark grey mantle and pale whitish grey lower parts, yellow bars on the wings, and grey marks on cheeks and throat where _pinus_ has black. These two birds are exceeding distinct, and in addition their songs are quite unlike. _H. pinus_ ranges through the eastern United States up to Connecticut and Iowa. _H. chrysoptera_ is a northern form extending down to Connecticut and New Jersey. Both are migrants.

In these two States, where the two types overlap, certain forms have been repeatedly found which have been described as two distinct species, _Lawrencei_ and _leucobronchialis_. Dr. L. B. Bishop and Mr. Brewster showed me two long series of _Helminthophila_ containing various intergrades between the four named kinds, and details regarding these may be found in Chapman's _North American Warblers_ and in Dr. Bishop's paper in Auk, 1905, XXII. Though the characters evidently break up to some extent, the series can be represented as due to recombinations of definite factors more easily than the others which I have described. The differentiating characters are:

_Pinus_

1. Mantle and lower parts _yellow_ (Y^1). 2. Wing-bars _white_ (y^2). 3. Cheek and throat _not black_ (b).

_Chrysoptera_

1. Mantle and lower parts _grey_ (y^1). 2. Wing-bars _yellow_ (Y^2). 3. Cheek and throat _black_ (B).

The grey pigment of the mantle is common to both, but is masked by the yellow in _pinus_, the net result being an olive-green.[10]

I am much indebted to Dr. F. M. Chapman for the loan of the coloured plate in which these distinctions are shown. It first appeared in his book, _North American Warblers_.

We cannot tell whether _yellow_ or _not-yellow_ is due to the presence of a factor, but we may suppose that one or other gives the special colour to the parts. The black of character 3 is no doubt a dominant. Thus _pinus_ becomes Y^{1}y^{2}b and _chrysoptera_ in y^{1}Y^{2}B. The _Lawrencei_ which has the underparts _yellow_, wing-bars _white_, and _black_ patches is Y^{1}y^{2}B and _leucobronchialis_ which has mantle and underparts _not-yellow_, wing-bars _yellow_ and _no black patches_ is y^{1}Y^{2}b. This representation, it should be clearly understood, is tentative and approximate only. The characters are not really sharp, for there is much grading; but allowing for the effects of heterozygosis and for some actual breaking-up of factors I believe it gives a fairly correct view of the case. In particular we can see how it meets the difficulty which Chapman felt in accepting _leucobronchialis_ as in any sense derived from _pinus_ which has a yellow breast, and _chrysoptera_ which has a black throat, seeing that _leucobronchialis_ has neither. We now recognize at once that this form could be produced by ordinary re-combination of the absence of Y^{1} with the absence of B.

I note also with great interest that the modern observers agree that the so-called hybrids may have the song either of the one species, or of the other, or a song intermediate between the two. It may also be added that these two types have several times been seen, in the breeding season, paired with each other or with one of the other combinations.

Allen[11] has described another excellent American example, the Tits of the group _Baeolophus bicolor-atricristatus_. The form _bicolor_ belongs to the eastern States and ranges from the Atlantic coast to the Great Plains, and _atricristatus_, of east Mexico, extends from Vera Cruz to central Texas. In southern and central Texas the breeding ranges adjoin, and in this country various intermediates occur. The chief types differ in two main points.

_B. bicolor_

Forehead varies from deep _black_ to dull black, suffused with rusty brown.

Crown and crest _grey_, slightly darker than the back.

_B. atricristatus_

Forehead _white_ to buffish white.

Crown and crest _black_, abruptly contrasting with the back.

The intergrades between the two have, as usual, received specific names. A detailed description is given by Allen, from which it appears that the gradation is very complete. In one case a series of 16 adults were all intermediates. It is not stated whether the collector took these at random, but from the local lists it is clear that the types are found not far away from the place where the intergrades were shot.

Another very striking case is that of the Tanagers, of the genus _Rhamphocoelus_. In this group there are several local forms which are related to each other in remarkable ways. The forms known as _passerinii_ and _icteronotus_ exhibit the clearest phenomena of intergradation. The species _passerinii_ has a brilliant scarlet and black male, and it inhabits Honduras and Nicaragua. Proceeding southwards along the isthmus we find next _costaricensis_ which has a male like that of _passerinii_ (but a female with more orange than the olive-grey female of _passerinii_). Next we come to Panama which is occupied by _icteronotus_, sharply distinguished from _passerinii_ by the fact that the _scarlet is replaced by lemon-yellow_. This same _icteronotus_ occurs again as a pure type in Ecuador and many other parts of South America; but Colombia, _between Panama and Ecuador_, contains scarlets like _passerinii_, yellows like _icteronotus_, and various intergrades of several shades of orange. The _passerinii_ males from Nicaragua are indistinguishable from those of Colombia, and the _icteronotus_ of Ecuador are the same as those in Panama. The orange intergrades, doubtless heterozygous forms, though collected at the same locality (Medellin in Colombia) as several pure yellows and pure scarlets, are in the British Museum series sorted out as a separate species under the name _chrysonotus_! Complications are introduced by the relations of these forms to another named type, _flammigerus_, but we may for our purpose leave that out of consideration, and say that the order of geographical sequence from Honduras to Ecuador is (1) scarlet, (2) yellow, (3) mixture of types, scarlet, yellow, orange, (4)yellow.

Similar examples exist in the birds of the old world, but I do not know of any that have been studied so fully as those of America. The best known is that of the two Rollers, _Coracias indicus_ which spreads from Asia Minor through Persia, Baluchistan, the Indian Peninsula and Ceylon, and _affinis_ which ranges from Nepal, through Assam, Tenasserim and the Indo-Chinese countries. The two types are very different and may be distinguished as follows:

_C. indicus_

_Mantle_ drab brown-chestnut. _Breast_ chestnut. _Throat_ purplish, streaked with white. _Upper tail-coverts_ indigo.

_C. affinis_

Dark olive-green. Dull purple brown. Purple, streaked with blue. Turquoise.

The wings are the same in both. In the provinces of Nepal, Sikhim, and Darjiling the two species coexist, with the result that intergrades have been frequently recorded. The line of intergradation extends to the coast, and birds showing various combinations of the two types from the Calcutta district exist in collections.[12] The case is interesting inasmuch as like that of _Quiscalus_ it shows a series of combinations of various metallic colours. Some of these are probably evoked by the development of pigment behind striations or other interferences already existing, but in the present state of knowledge it would be quite impossible to suggest what the actual factors producing these appearances may be.

There are, naturally, many other cases among birds which are suspected of being in reality comparable, but in most of them the evidence is still inadequate. Among Lepidoptera also there are a few of these; perhaps the most striking is that of _Basilarchia "proserpina."_[13] The genus is well known to European collectors under the name _Limenitis_, of which we in England have one species, _L. sibylla_, the "White Admiral." A species very like _sibylla_ in general appearance is common in the northern parts of the United States, ranging through Canada and Northern New England, but rarely south of Boston. This species has the conspicuous white bands across both wings like our _sibylla_.

There is also a more Southern type known as _astyanax_, which is very different in its appearance, being without the white bands and having a broad irroration of blue scales on the posterior border of the hind wings. The two are so distinct that one would not be tempted to suspect any very close relation between them. In its distribution _astyanax_ is described by Field as replacing arthemis south of latitude 42°. About Boston it is much more common than _arthemis_.

The two forms encroach but little on each other's territory, but where they do coexist, a third form, known as _proserpina_, is found which is almost intermediate, with the white bands much reduced. There is now no doubt that this _proserpina_ is a heterozygous form, resulting from a combination of the characters of _arthemis_ and _astyanax_. Field succeeded in rearing a brood of 16 from a _proserpina_ mother caught wild which laid 31 eggs, and of these, nine (five males, four females) resembled the mother, being _proserpina_, and seven (four males, three females) were _arthemis_. There can be no question therefore that the mother had been fertilised by a male _arthemis_ and that _no-white-band_ is a factor partially dominant over the _white band_. Another point of interest which Field observed was that the _proserpina_ female refused to lay on birch, poplar or willow, but accepted wild cherry (_Prunus serotina_) a species on which _astyanax_ can live, though that tree is not known to be eaten by _arthemis_. Incidentally also the observations show that sterility cannot be supposed to be the bar which maintains the distinctness of _arthemis_ and _astyanax_.

In this connection _Papilio oregonia_ and _bairdii_ should be mentioned.[14] _P. oregonia_ is one of the numerous forms like _machaon_, but rather paler. It is a northern insect, inhabiting British Colombia east of the Cascade Range, and reaching to Colorado. _P. bairdii_ is a much darker butterfly, representing the _asterias_ group of the genus _Papilio_. Like _asterias_ it has the abdomen spotted at the sides, not banded as in the _machaon_ group. It belongs to Arizona and Utah extending into Colorado. From Colorado the form _brucei_ is described, more or less intermediate, like _bairdii_ but with the abdomen banded as in _oregonia_. W. H. Edwards records the results of rearing the offspring of the _bairdii_-like and of the _oregonia_-like mothers. Each was found able to have offspring of both kinds, that is to say, _bairdii_ females gave both forms, and _oregonia_ females gave both forms. It is not possible to say which is dominant, since the fathers were unknown. On general grounds one may expect that the _bairdii_ form will be found to dominate, but this is quite doubtful.

From this particular discussion I omit reference to those examples in which the permanently established types are obviously associated with special conditions of life. Where considerable climatic differences exist between localities, or when we pass from South to North, or from the plains into Alpine levels we often find that in correspondence with the change of climate there is a change in the characteristics of a species common to both. When I say "species" in such a connection I am obviously using the term in the inclusive sense. Some would prefer to say that in the two sets of conditions two _representative species_ exist. Whichever expression be preferred it is plain that such examples present another phase of the problem we have been just considering, and in them also we have an opportunity of observing the consequences of the overlap of two closely related types, but there are advantages in considering them separately. In the examples hitherto given, with the possible exception of the Papilios,[15] the two fixed types severally range over so extensive a region that it may fairly be supposed that in the different parts they are subject to considerable diversities of climate. There is no outstanding difference that we know distinguishing the habitats of the two forms; but in comparing Alpine with Lowland forms, or essentially northern with essentially southern forms we do know an external circumstance, temperature, that may reasonably be supposed to have an influence, direct or indirect, on the population.

FOOTNOTES:

[1] J. A. Allen, _The North American Species of the Genus Colaptes, Considered with Special Reference to the Relationships of C. auratus and C. cafer_. Bull. Am. Mus. Nat. Hist., IV, 1892.

[2] For a case in which a red-headed female × a black-headed male gave a black-headed female and a red-headed male, see _Avian Mag._, N. S., IV, pp. 49 and 329

[3] The other variations of this bird are also interesting and important. The normal male has a red head and a red throat. The female has a red head and a white throat, but varieties of the female are known with a black head, thus again illustrating the change from black to red. It should be noted that this is not a mere retention of a juvenile character, but, as the birds mature, the red feathers come up, or as an exception, the black. There is also a western species, _ruber_, in which both sexes have a great extension of red, and are alike. The male of _nuchalis_ intergrades with this type, but the female does not.

[4] Dr. W. Brewster, for example, has a remarkable specimen of the Teal (_Nettion carolinense_) with a white collar strongly developed at the front and sides of the neck, in a place where the normal has no such mark.

[5] This variety is spoken of as the Ringed Guillemot and is sometimes regarded as a distinct species to which the name _ringvia_ was given by Brünnich. In support of this view Dr. William Brewster, to whom I am indebted for much assistance in regard to the variation of birds, called my attention to observations of his own and also of Maynard's, that the ringed birds were sometimes mated together, though in a small minority (see Brewster, _Proc. Boston Soc. N. H._, XXII, 1883, p. 410). It would however be possible to produce many instances of varieties mated together though surrounded by a typical population (_e. g._, two varying Blackbirds, _Zoologist_, p. 2765; two varying Nightjars, _ibid._, p. 5278). I am inclined to believe that in nature matings between brothers and sisters are frequent in many species of animals, and that the production of sporadically varying colonies is thus greatly assisted.

[6] The Sap-suckers feed on trees and somewhat resemble our Spotted Woodpeckers in general appearance. _Colaptes_ feeds on the ground and corresponds perhaps rather with the European Green Woodpecker.

[7] For an introduction to this example I am indebted to Mr. W. D. Miller of the American Museum of Natural History. Some account of the facts is given by Baird, Brewer, and Ridgway (_A Hist. of N. Amer. Birds_. 1874, II, pp. 540, 544, etc.). _S. varius_ occupies the whole country in suitable places from the Atlantic to the eastern slopes of the Rockies, and all Mexico to Guatemala. _S. nuchalis_ was first known from the Southern Rockies only, but many were afterwards taken in Utah. _S. ruber_ is restricted to the Pacific coast. In Ridgway's opinion all three are geographical forms of one species. In _ruber_ the sexes are alike having both a great extension of the red in the throat, and a red crescent. The male of _nuchalis_ grades to the _ruber_ form, but the female does not. This female has some red in the throat like the male of _varius_, whereas the female of _varius_ has a whitish throat.

[8] Not only vertebrates but the marine Crustacea and Mollusca illustrate this curious "principle" of variation, as Canon Norman formerly pointed out to me with abundant illustrations. There are of course cases to the contrary also.

[9] Chapman, F. M., _Bull. Amer. Mus._, IV, 1892, p. 1; see also Ridgway, _Birds of North and Middle America_, 1902, Part II, p. 214.

[10] It would aid greatly in factorial analysis if the descriptive term "green" could be avoided in application to cases where the green effect is due only to a mixture of black and yellow pigments. The absence of yellow is the sole difference between the mantle and underparts of _pinus_ and _chrysoptera_.

[11] _Bull. Amer. Mus. Nat. Hist._, XXIII, 1907, p. 467.

[12] References on this subject will be found in _Brit. Mus. Cat. Birds_, XVII, p. 13.

[13] For these facts I am indebted to Mr. W. L. W. Field, who has lately published an account of his observations and experiments. See especially, _Psyche_, 1910, XVII, No. 3, where full references to previous publications are given.

[14] For the facts and further references see W. H. Edwards, _Butterflies of N. America_, 2d series, Papilio VII and X; 3d series, 1897, Papilio IV, _Can. Entom._, 1895, XXVII, p. 239.

[15] I think this case is fairly included because the _machaon_ type is so widespread that it cannot be regarded as a product of a Northern climate, nor can _asterias_ be claimed as especially a warm country form, seeing that _brevicauda_, which is scarcely distinguishable from _asterias_, inhabits Newfoundland (having a curious phase there in which the yellow is largely replaced by red).