by the splitting of one old one, to parthenogenesis, or reproduction from unfertilized ova, without finding any important gap in the series, and we may safely conclude that all these forms of reproduction are fundamentally alike.

So far as regards the physical side of the problem of heredity, the only essential difference between asexual reproduction and sexual reproduction is the absence of fertilization or union with a male cell in the one case, and its occurrence in the other case.

It is therefore extremely important to compare the two processes, in order to discover whether this physical difference is accompanied by any difference in the result. In the one case we have heredity with the male factor omitted, and in the other we have heredity with a male factor, and if there is any constant difference in the result, we may safely attribute it to this factor.

In making this comparison we are almost compelled to restrict ourselves to plants, for although asexual reproduction is not at all unusual in animals, it is restricted, with one exception, to animals which are not domesticated or reared by man, and we therefore know too little about the minute details of their life to make use of them for our purpose. The number of plants which have been. cultivated and carefully observed and studied by man is very great, and as most of them multiply asexually by budding, as well as by fertilized seeds, we here have abundant material for comparative study, and it is well established by hundreds of thousands of observations that the presence or absence of the influence of the male element does have an influence upon the result of the reproductive process, and that this result is exactly what our view of the nature of the process would lead us to expect. Plants produced from fertilized seeds differ from those

produced from buds only in their greater tendency to vary. Bud variations do occur, but they are very unusual, while more or less variation in seedling plants is almost universal.

As we suppose that any cell may, when excited by unfavorable conditions, throw off gemmules, the gemmules may find their way, by a sort of accident, to growing buds, and thus cause variation. We should therefore expect bud variation to occur occasionally, but very much less frequently than variation in seedlings.

This is so well known to be the case that many authors have held that there can be no variation without sexual union. Darwin has shown, however, by a long list of instances of bud variation in plants, that this is not absolutely true, and the weight of his authority has led to the almost universal acceptance of his conclusion that there is no essential difference between asexual and sexual heredity. I shall discuss this conclusion at length in another place, as I believe that the facts demand an interpretation which is somewhat different from the one which Darwin furnishes. At present I simply wish to call attention to the fact that all authorities agree that variation is almost infinitely more common in sexual than it is in asexual offspring.

Asexual multiplication in animals is restricted to the lower forms which are of little use to man, and as these forms have not been domesticated and carefully observed, our knowledge of the variability of organisms produced asexually is almost entirely derived from the study of plants.

The only instance in domesticated animals of anything like asexual reproduction is the parthenogenetic reproduction of bees, and it is therefore interesting to note

that the hive-bee is the least variable of all domesticated animals (Darwin, Variation, Vol. ii. p. 307).

Darwin says (Variation, Vol. i. p. 360) that he procured a hive full of dead bees from Jamaica, where they have long been naturalized, and on carefully comparing them under the microscope with his own bees, could not detect a trace of difference.

With plants it is well known to all cultivators that forms which are highly variable as seedlings can be kept perfectly true by asexual propagation, and we have Darwin's authority (Variation, Vol. ii. r. 307, and Vol. i. p. 429) for the statement that while hardly a single plant can be named which has long been cultivated and propagated by seed that is not highly variable, the total number of instances of bud variation is as nothing in comparison with seminal varieties.

This contrast is the more remarkable when we recollect that in most of our cultivated plants the number of buds which develop is thousands of times greater than the number of seeds which give rise to plants. It is clear that if the chance of variation were the same in both cases the number of bud variations would be thousands of times greater than the number of seedling variations. If there were thousands of chances of seedling variation for one chance of bud variation, the number of bud varieties would still be equal to the number of seedling varieties.

The fact that with all this probability in their favor, bud varieties are very rare as compared with seedling varieties, shows that the chance of bud variation is almost infinitely small as compared with the chance of seedling variation.

While we cannot deny that variation may sometimes occur in organisms produced asexually, I think we are

justified in giving great emphasis to the law that variability is almost exclusively the characteristic of organisms produced from fertilized ova.

Darwin says (Variation, Vol. ii. pp. 351 and 377), "When we reflect on the millions of buds which many trees have produced before some one bud has varied, we are lost in wonder what the precise cause of each variation can be." "Habit, however much prolonged, rarely produces any effect on a plant propagated by buds: it apparently acts only through successive seminal generations."

The curious history of the naturalization of the orange in Italy, quoted by Darwin on the authority of Gallesio (Theoria della Riproduzione Veg, 1816, p. 125), is very interesting in this connection. During many centuries the sweet orange was propagated exclusively by grafts, and so often suffered from frost that it required protection. After the severe frost of 1709, and more especially after that of 1763, so many trees were destroyed that seedlings from the sweet orange were raised, and to the surprise of the inhabitants their fruit was found to be sweet. The trees thus raised were larger, more productive and hardier than the former kinds, and seedlings were now constantly raised.

Hence Gallesio concludes that much more was effected for the naturalization of the orange in Italy by the accidental production of new kinds from seeds during a period of about sixty years than had been effected by grafting old varieties during many ages.

It is hardly necessary to give other illustrations of this law, for no one with any knowledge of the subject will be inclined to question it. It is strange that its significance has been overlooked, but this is probably due to the failure of students of the subject to perceive that it

is possible to believe that the transmission of variability is the peculiar function of the male cell, and also to acknowledge that variation may occasionally occur without its influence.

Our theory that variation is caused by the transmission of gemmules, and that there is no especial arrangement for their transmission to buds or to unfertilized eggs, while there is a special adaptation which has been slowly evolved during the evolution of sex for transmitting them to fertilized eggs, gives us a simple explanation of the fact that while bud variation is perfectly possible, it is extremely rare as compared with the variability of sexual offspring.

Darwin has been led, through the study of variability, to a conclusion which is very much like the explanation which is here presented. He says (Variation, Vol. ii. p. 325) that "we may infer from the occurrence of bud variation that the affection of the female element through external conditions may induce variability, for a bud seems to be the analogue of an ovule. But the male element is apparently much oftener affected by changed conditions, at least in a visible manner, than the female element or ovule.'

Bud variation is much more frequent in cultivated plants than it is in wild ones. Very few instances have ever been observed in plants growing wild or under strictly natural conditions, and Darwin states that "bud variation is most common in plants which have been highly cultivated for a long time."

The adjustment between a cultivated organism and its artificial or unnatural environment must, in most cases, be less perfect than that which has been slowly established between a wild organism and its natural environment. We should, therefore, expect domesticated and