ization of the gasteropod Physa, and have reached exactly the same result as that obtained in the echinoderms. Here also the egg-centre degenerates, and its place is taken by a centrosome brought in by the spermatozoon and giving rise to a sperm-amphiaster, which persists as the cleavage-amphiaster (Fig. 64). A strong presumption is thus created that Conklin was in error; and if this be the case, the last positive evidence of a conjugation of centrosomes in the animal egg disappears.' In view of this result we may well hesitate to accept Guignard's conclusions in the case of flowering plants. The figures of this author show in the clearest manner four centrosomes lying in the neighbourhood of the apposed germ-nuclei (Fig. 8o) ; but the conjugation of these centrosomes was an inference, not an observed fact, and has not been confirmed by any subsequent observer. Until such confirmation is forthcoming we must receive Guignard's results with The second view, based mainly upon the case of Myzostoma as described by Wheeler ('95), apparently rests on strong evidence, though its force cannot be exactly estimated until a more detailed account has been published. In this case no sperm-aster can be seen at any period, with which is correlated the fact that no middlepiece can be made out in the spermatozoon. The egg-centrosome, on the other hand, is stated to persist after the formation of the second polar body, to become double at a very early period, and to give rise directly to the cleavage-amphiaster (Fig. 78). I can find no ground in Professor Wheeler's paper to doubt the accuracy of his conclusions. Nevertheless, an isolated case, which stands in contradiction to all that is known of other forms, must rest on irrefragable evidence in order to command acceptance. Since, moreover, the case involves the whole theory of fertilization based on other animals (cf. p. 141), it must, I think, await further investigation.
I Richard Hertwig has, however, recently published a very interesting observation which indicates that we may not yet have fully fathomed the facts in the case of echinoderms. If unfertilized echinoderm-eggs, after formation of the polar-bodies, lie for many hours in water or be treated with dilute poisons (strychnine), they may form a more or less perfectly developed amphiaster, and the nucleus may even make an abortive attempt at division. No centrosomes, however, could be discovered, even by the most approved methods. This remarkable phenomenon is probably of the same nature as the formation of artificial asters observed by Morgan (p. 226), but its meaning is not clear.
2 Van der Stricht, in a recent paper on Amphioxus ('95), is inclined to believe that a fusion between the egg-centre and the sperm-centre occurs; but the evidence is very incomplete, and a comparison with the case of Physa indicates that his conclusion cannot be sustained. The same criticism applies to the earlier work of Blanc (91, '93) on the trout's egg.