and xxviii.) is readily accessible, but it will not be out of place to quote it before entering upon its critical dis cussion. He says: "In the previous chapters large classes of facts, such as those bearing on bud-variation, the various forms of inheritance, the causes and laws of variation, have been discussed, and it is obvious that these subjects, as well as the several modes of reproduction, stand in some sort of relation to each other. I have been led, or rather forced, to form a view, which to a certain extent connects these facts by a tangible method. Every one would wish to explain to himself, even in an imperfect manner, how it is possible for a character possessed by some remote ancestor suddenly to reappear in the offspring; how the effects of increased use or disuse of a limb can be transmitted to the child; how the male sexual element can act not solely on the ovule, but occasionally on the mother form; how a limb can be reproduced on the exact line of amputation, with neither too much nor too little added; how the various forms of reproduction are connected, and so forth. I am aware that my view is merely a provisional hypothesis or speculation, but until a better one be advanced it may be serviceable by bringing together a multitude of facts which are at present left disconnected by any efficient cause. As Whewell, the historian of the inductive sciences, remarks, hypotheses may often be of service to science, when they involve a certain portion of incompleteness or even of error. "Under this point of view I venture to advance the hypothesis of pangenesis, which implies that the whole organization, in the sense of every separate atom or unit, reproduces itself. Hence ovules and pollen grains-the fertilized seed or egg, as well as buds-include and con sist of a multitude of germs thrown off from each separate atom of the organism." From the extract we see that the hypothesis is an attempt to show that all the phenomena of generation and development, including those of variation as well as those of heredity, depend upon the fact that each structural unit of the body is the direct offspring of a similar unit in the body of a parent or of a more remote ancestor. The cells of the body of one of the higher organisms are not only morphologically but actually independent individuals, reproducing themselves directly in the next generation: and the germ of such an organism is in reality an aggregate of these cell-germs. Stated more at length, the hypothesis is as follows: "I assume that cells, before their conversion into 'form material,' throw off minute granules or atoms, which circulate freely throughout the system, and when supplied with proper nutriment, multiply by self-division, subsequently becoming developed into cells like those from which they were derived. These granules, for the sake of distinctness, may be called gemmules. They are supposed to be transmitted from the parent to the offspring, and are generally developed in the generation which immediately succeeds, but are often transmitted in a dormant state during many generations and are then developed. Their development is supposed to depend on their union with other partially developed cells or gemmules, which precede them in the regular order of growth. Why I use the term union will be seen when we discuss the direct action of pollen on the tissues of the mother plant. "Gemmules are supposed to be thrown off by every cell or unit not only during the adult state but during all stages of development. Lastly I assume that gemmules . in their dormant state have a mutual affinity for each other, leading to their aggregation either into buds or into the sexual elements. Hence, speaking strictly, it is not the reproductive elements nor the buds which generate new organisms, but the cells themselves throughout the body. These assumptions constitute the provisional hypothesis of pangenesis." Darwin's gemmules are, of course, entirely.imaginary, that is, a belief in their existence does not rest upon direct observation. We cannot deny that the hypothesis. furnishes an explanation of most of the phenomena which he attempts to interpret by it, although it seems possible that there may be a simpler explanation. If the existence of the gemmules were proven we could understand not only the wonderful facts of ordinary inheritance by sexual reproduction, but the various forms of asexual reproduction as well. We should have a simple explanation of the manner in which the characteristics of a remote ancestor may suddenly reappear after they have been dormant for many generations. We should understand how the embryological history of a species may become simplified by the omission of larval forms or appendages. In a word, nearly all the phenomena of heredity admit of explanation by the hypothesis, and those who have criticised it have not usually attempted to show that it conflicts with fact, but have simply objected to it as a purely imaginary explanation. It is urged that the transmission of all the characteristics which we know to be inherited from near and remote ancestors demands that the number of gemmules should be almost unlimited and practically infinite; that not only are the gemmules imaginary, but that the aggregation of such numbers in masses as small as the reproductive elements requires that they shall be of inconceivable minuteness, and that nature furnishes no analogy for attributing to such small particles the vital properties which we know only in bodies which are comparatively gigantic. It is also urged that the gemmules must be endowed with entirely imaginary and wonderfully specialized elective affinities, in virtue of which each develops only at the proper time and place. In order to account for the manner in which the characteristics of each parent are mingled in the child we must regard each individual as the product of a struggle for existence among the gemmules, resulting in the selection and development of the fittest. The formation of several individuals asexually by budding from a parent stock demands that the gemmules themselves must be capable of multiplication, and that they must have the power to transmit their properties to their offspring. To explain alternation of generations we must suppose that the embryo receives several complete sets of gemmules, which are not duplicates, and it is almost impossible to follow out, in thought, the complicated relations which must exist between the gemmules of the egg-embryo of such an organism as a Siphonophore. These and similar objections may be fairly urged, and while their great weight is obvious, we must not attach undue importance to them, for they do not show that the hypothesis conflicts with any known law or observed fact, and the great drafts made upon the imagination should not, alone, prevent its provisional acceptance so long as we have no simple explanation of the phenomena, for difficulty in imagining the details of an hypothesis is a purely subjective matter, which varies with the age and with the individual. Galton's Experiments. Besides these theoretical objections, we have the experimental disproof furnished by Galton. In order to test the hypothesis this experimenter selected the silvergray rabbit-a variety which has, in itself, little tendency to vary, although it readily crosses with other varieties, and breeding freely with them gives birth to hybrid offspring. Into the bodies of eighteen of these silver-gray rabbits he transfused the blood of other varieties, in some cases replacing one half of the blood. From the eighteen rabbits thus operated upon eighty-six young were produced, and in no case did the offspring exhibit any of the characteristics of the variety from which the blood was taken, but all of the eighty-six were pure silver gray. From these experiments Galton concludes that "the doctrine of pangenesis, pure and simple, is incorrect;" and I think we must agree with him that this conclusion is justified by the results which he reached, although I hope to show that it is possible to restate the hypothesis in a form which is so modified as to escape this objection. The Sexual Elements Perform Different Functions in There is another objection which seems to me to be of almost equal weight, but which has never, so far as I am aware, been pointed out. The early writers upon heredity attributed certain functions to the male cell and others to the ovum; but we now know that their means of observation were so inadequate, and their knowledge so limited, that their conclusions were of little value, and that both ovists and spermists were wide of the mark. The fact that they erroneously attrib |