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).