EFFECT OF NATURAL SELECTION.

91

which, although sexually mature, is evidently only a dwarf form of the deep-sea E. Flemingii*.

There are two ways in which variation in the period at which reproduction commences may, and probably has given rise to very considerable modifications-first, by the agency of natural selection, and, secondly, by direct variation produced by changed conditions.

It is perfectly clear that each species will derive advantage from commencing to multiply at as early a period as is consistent with the maintenance of characters adapted to the mode of life and the economy of the species. Hence the great majority of animals tend to produce young before they are perfectly mature. In the higher forms of life no considerable modification is likely to arise from this condition, as more individuals are always produced after maturity than before; so that even in cases in which any considerable advantage arose from such an early brood, it would not probably materially affect the species. In the lower forms, however, in which a large number of individuals are produced at one time, the case is very different.. The early breeding of a species may be of very material advantage; the later broods being cut off, for instance, by the approach of winter or dearth of food, the earlier alone would survive; and thus in

*W. Thomson, The Depths of the Sea,' p. 117.

92

DIRECT ACTION OF CONDITIONS.

course of time the adult characters would be liable to be lost. In the species of Hemiptera already cited, the extremely rare occurrence of adult characters is thus accounted for.

Instances of the direct action of conditions are seen in highly fed domestic animals, which become mature at an early period of life. In the human race it is well known that girls arrive at womanhood at an earlier period in hot than in cold countries, perhaps from the custom of early marriages. Amongst the lower forms of life such changes in the period of maturity, although their effects are not very appreciable in the higher forms, give rise to great changes of form and structure.

Changes of period in the lower animals would, as Mr. Cope has suggested, produce very remarkable changes of form. The immature sexuality of the larvæ of insects, if the mature forms were lost, would, for instance, produce a new division of the Annulosa. Perhaps "we see the effect of accelerated development of the ovaries in the viviparous agamic generation of Cecidomyian larvæ. Dr. Leuckart's observations* leave little doubt that the germ-stocks are actually modified ovaries, and that the development of the new larvæ within the body of the mother is the result of the non-development of the rest of the sexual

* Ann. & Mag. Nat. Hist. 1866, 3rd series, vol. xvii.

EARLY VARIATION IN DEVELOPMENT.

93

apparatus. The agamic nature of the process is no objection to this view, as we know that parthenogenesis is by no means uncommon amongst perfect insects even when they are ready to produce young in the ordinary way." In the agamic reproduction of Aphides we observe the same tendency to the non-development of the accessory organs of generation. The pupa of Chironomus exhibits a similar phenomenon, as Grimm has shown, thus uniting the parthenogenesis of adult and larval forms *.

The developmental process is probably subject to variations from its very commencement. This is seen in some kinds of malformations; arrest of development, union of the fingers or toes, and supernumerary digits arise from slight deviations occurring at very early periods of life t. These may be extremely slight in their commencement, and yet their effects may be very considerable. As a rule it appears that the earlier any variation occurs, however slight at first, the graver its effects on the adult form. Those who have worked most carefully at the subject of teratology admit that no definite line can be drawn between ordinary varieties and the lesser forms of malformation. The

* Darwin, 'Origin of Species,' p. 387.

+ Teratological Catalogue of Specimens in the Hunterian Museum,' by the Author.

aortic arches present considerable variation in their mode and order of obliteration; and this gives rise to extreme variety in the distribution of the vessels of the adult. Von Baer has shown that the disposition of the viscera is liable to be entirely reversed by a very slight variation early in the process of development; and it is exceedingly difficult to see the limit to the change which might occur in a very few generations by very slight deviations from the usual process of development in its earlier stages, if these became useful to the individual. In domesticated animals and in Man, all known deviations of development are so considerable that they could hardly be advantageous. This is probably owing to the great plasticity of the ovum and its susceptibility to disturbing causes.

On the other hand, a comparison of the rudimentary toes of some species of deer with digits undergoing abnormal atrophy in certain malformations will show at once the same process manifested in both. Fusion of bones is occasionally observed from arrest in the development of some of the histological tissues between them; and similar fusion occurs normally in nature. Webbing of the digits and the formation of a patagium or flying membrane occur both as normal and abnormal conditions. As such characters appear suddenly from early deviations of development

PERIOD OF DEVELOPMENT.

95

in abnormal conditions, it is not difficult to believe that they arise suddenly and are selected when they are useful. Although it would be unwise with our present knowledge to assert positively that such changes have really given rise to new and permanent forms, nevertheless the hypothesis explains satisfactorily the occurrence of webbed toes and flying membranes in very different highly developed animals.

There can be no doubt that variations are most frequent during the later periods of development, and that those which occur in the earlier stages are usually disadvantageous, as they give rise to very considerable modification of structure; useful variations would undoubtedly usually be so slight in the young embryo that they would pass unperceived. The apparent invariability of embryos is probably due to this; and hence they afford the most important evidence of the line of descent by which any form has been perfected.

Another point of considerable interest in this connexion is that organs are developed at an earlier period in some types than in others. The common duck, for instance, is hatched with exceedingly rudimentary wings when compared with the chick; and this difference is even more marked about three or four weeks later, when the wings of the duck are still, like those of the penguin, mere