the dead and the living were carefully measured, with the result that the survivors always had a smaller frontal breadth, agreeing with the change that had taken place during several years' exposure to a much less deleterious amount of the clay. In order to test the result more accurately; some of the finest sediment was collected from the very spots on the shore from which the large number of crabs first experimented with were obtained, and this gave exactly the same result, indicating a selective action by this kind of impure water leading to a change in the proportions of these crabs, which, if continued at anything like the same rate even for a century, would so alter the shape of the animals as to produce what would certainly be described as a new species.

Experiments of this kind are still going on; every precaution is being taken to secure trustworthy results; and there seems little doubt that we have here a means of seeing natural selection actually at work.

Selective Elimination of Sparrows in America.—Another experiment has recently been made in America with equally interesting results.

In February 1898, after a very severe storm of snow, rain, and sleet, 136 English sparrows were found benumbed and were brought to the Anatomical Laboratory of the Brown University, Providence, Rhode Island. Under identical conditions, 72 of these birds recovered, while 64 perished; and the happy thought occurred to Professor H. C. Bumpus of trying to ascertain whether there were any physical peculiarities distinguishing the one group from the other -in other words, Was storm a selective agency which eliminated those less fit to withstand it? Mr. Bumpus accordingly made a rigid examination both of the dead and the living birds, which were fortunately nearly equal in numbers, with the following results.

1. As to sex; it was found that females suffered more than males. Of the former 21 lived while 28 died, while of the latter 51 lived but only 36 died. It is therefore clear that, as might be expected, the female birds had a constitution less able to withstand severities of climate than the males.

2. Size. Here the comparison is made between adult and young male birds separately, and of females with females. In all three of these groups those which perished were on the average of greater length than those which survived. Of the 35 adult males which survived less than half reached 160 millimetres, while of 24 which perished all but two were of that size and upwards. Of the females, less than one-fourth of those that survived reached the same size, while nearly half of those that perished did so. In the young males there was a nearly similar disproportion. There is clearly no evi

dence of chance here. The very same kind of elimination occurred in all three groups, showing that, on the average, large birds (of the same species, age, and sex) are less able to withstand wind and cold than smaller ones.

3. Weight. This character gives the same result as that of length, the survivors being on the average lighter than those that died, to the large amount of about one twenty-fifth of the weight.

4. Length of the Sternum. This character gives a rather unexpected result, those birds which survived having a decidedly longer sternum than those which perished. The difference is about 013 of the length; but as the smaller birds on the whole survived, they evidently had their sternums proportionally very long. As the sternum indicates the size of the pectoral muscle, this would show that the surviving birds had a stronger flight, and also that there was a larger surface of this muscle in proportion to that of the whole body, giving less exposure of the visceral organs.

The result of the whole inquiry is very satisfactory, since it clearly demonstrates an eliminating force due to conditions to which all birds, as well as other animals, are exposed. One of the most constantly adduced objections to the theory of natural selection is, that the comparatively small individual variations on which Darwinians rely as furnishing materials for the origin of new species by adaptation to new conditions, are not of sufficient importance and magnitude to determine life or death- -are not, as Professor Lloyd Morgan terms it, of "survival value." But here we have a striking proof that they are so. For the slight variations in size and proportions shown by the birds here measured are of the exact kind that occur always and everywhere among the individuals of all common species in a state of nature, and they are here shown to be of "survival value."

Taking these observations in connection with those on crabs, showing a progressive change of shape due to new and somewhat injurious conditions, it can no longer be said that natural selection is a pure theory, and has never been shown to be actually effective in nature. The observations here briefly described may not be sufficient, and will no doubt be greatly extended in the future; but so far as they go they exactly conform to what, if the theory of natural selection be true, must be of constant occurrence in nature, and prove that survival of the fittest is not only a theory but an observed fact.

The above is an abstract of a Lecture, a copy of which has been sent me by the author.

CHAPTER VI

DIFFICULTIES AND OBJECTIONS

Difficulty as to smallness of variations-As to the right variations occurring when required-The beginnings of important organs-The mammary glands-The eyes of flatfish-Origin of the eye-Useless or non-adaptive characters-Recent extension of the region of utility in plants-The same in animals-Uses of tails-Of the horns of deerOf the scale-ornamentation of reptiles-Instability of non-adaptive characters-Delboeuf's law-No " 'specific" character proved to be useless-The swamping effects of intercrossing-Isolation as preventing intercrossing-Gulick on the effects of isolation-Cases in which isolation is ineffective.

IN the present chapter I propose to discuss the more obvious and often repeated objections to Darwin's theory, and to show how far they affect its character as a true and sufficient explanation of the origin of species. The more recondite difficulties, affecting such fundamental questions as the causes and laws of variability, will be left for a future chapter, after we have become better acquainted with the applications of the theory to the more important adaptations and correlations of animal and plant life.

One of the earliest and most often repeated objections was, that it was difficult "to imagine a reason why variations tending in an infinitesimal degree in any special direction should be preserved," or to believe that the complex adaptation of living organisms could have been produced "by infinitesimal beginnings." Now this term "infinitesimal," used by a wellknown early critic of the Origin of Species, was never made use of by Darwin himself, who spoke only of variations being "slight," and of the "small amount" of the variations that might be selected. Even in using these terms he undoubtedly afforded

grounds for the objection above made, that such small and slight variations could be of no real use, and would not determine the survival of the individuals possessing them. We have seen, however, in our third chapter, that even Darwin's terms were hardly justified; and that the variability of many important species is of considerable amount, and may very often be properly described as large. As this is found to be the case both in animals and plants, and in all their chief groups and subdivisions, and also to apply to all the separate parts and organs that have been compared, we must take it as proved that the average amount of variability presents no difficulty whatever in the way of the action of natural selection. It may be here mentioned that, up to the time of the preparation of the last edition of The Origin of Species, Darwin had not seen the work of Mr. J. A. Allen of Harvard University (then only just published), which gave us the first body of accurate comparisons and measurements demonstrating this large amount of variability. Since then evidence of this nature has been accumulating, and we are, therefore, now in a far better position to appreciate the facilities for natural selection, in this respect, than was Mr. Darwin himself.

Another objection of a similar nature is, that the chances are immensely against the right variation or combination of variations occurring just when required; and further, that no variation can be perpetuated that is not accompanied by several concomitant variations of dependent parts-greater length of a wing in a bird, for example, would be of little use if unaccompanied by increased volume or contractility of the muscles which move it. This objection seemed a very strong one so long as it was supposed that variations occurred singly and at considerable intervals; but it ceases to have any weight now we know that they occur simultaneously in various parts of the organism, and also in a large proportion of the individuals which make up the species. A considerable number of individuals will, therefore, every year possess the required combination of characters; and it may also be considered probable that when the two characters are such that they always act together, there will be such a correlation between them that they will frequently vary together. But there is another consideration that seems to show that this coincident