observation is even more important. He discovered that there are large and small eggs laid by the silk-worm moth, and that the caterpillars emerging from the large eggs are, in the great majority of cases (88 to 95 per cent.), females, while those from the smaller eggs give a corresponding majority of males (88 to 92 per cent.). It is therefore clear that the difference in size corresponds to a difference in the sex of the embryos, and that with sufficient care it would be possible to separate the two sorts of eggs so completely that all of one kind would be females and of the other males. A similar result has been obtained by Joseph in another moth, Ocneria dispar. Cuénot states that he has been able to verify completely this statement of Joseph.
How can we explain the apparent contradiction between the results of Landois, Treat and Gentry, and those of Brocadello, Joseph and Cuénot? It is probable that in all cases the facts recorded are correct. Cuénot suggests that in the lots of caterpillars that were poorly nourished there was a large mortality of the young females, so that of those surviving there was a larger percentage of males. If we apply this same view to the case in which abundant feeding gives rise to more females we shall have to assume that here a larger percentage of males are eliminated, but it is not at all evident why this should be the case. Cuénot points out another possible source of error; namely, that in selecting the caterpillars for the experiment the larger ones may have been picked out to be given an abundant diet and the smaller ones for a meager diet. If this had been done consciously, or unconsciously, the results would not be due to the quality of the food, because the young caterpillars that were large were already females (having come from larger eggs), and those that were small were already males (having come from small eggs). It is clear, therefore, that all the early experiments must be repeated and the precaution taken to note the number of caterpillars at the beginning and at the end of the experiment, and the sex of those that die must, if possible, be determined. Care must also be taken that no selection of large and of small individuals takes place. Since, however, it has been so clearly shown by Brocadello and by Cuénot that large eggs become females and small eggs males, it will be desirable in repeating the experiments to take this fact into account, and to attempt to discover if the potentialities of the large and of the small eggs can be changed by external conditions. Here we have a new field for experimental work that will yield results of great interest. The lines are now so definitely drawn, and it is clearly so important to settle this question on as many forms as possible, that it is much to be hoped that in the near future many workers will turn their attention to this important subject.
Cuénot's experiments on certain flies, belonging to four different genera, are of especial interest. In this group also it had been claimed
