742 



SCIENCE. 



[N. S. Vol. XVlll. No. 4G7 



SO that the expei'imental control of the feed- 

 ing has been necessary only during the 

 larval or 'sillrvvorm' stage. The larval 

 life is subdivided into five stages clearly 

 set off from one another by the intervening 

 moults, of which there are normally four, 

 and these substages have been useful when 

 an alteration of food conditions during a 

 sharply defined shorter time than the er^ 

 tire larval life was desirable. 



The change in quality of food has con- 

 sisted in a substitution of lettuce for the 

 silkworm's proverbial mulberry diet. The 

 change in quantity of food has consisted 

 in altering the amount of mulberry served 

 to the larvte, the control of which has been 

 secured as follows : It has been determined 

 through experience with normal larvae that 

 each will consume a certain amount of food 

 in a certain number of hours (increasing 

 in amount with the increasing age and size 

 of the larva), this amount representing the 

 optimum amount of food for the normal 

 individual and necessitating as many daily 

 meals as are required to keep any but the 

 moulting larva constantly supplied with 

 fresh food. This amount determined, a 

 tolerably definite small proportion of the 

 optimum amount has been allotted the in- 

 dividuals which were sentenced to short 

 rations, which, roughly speaking, might be 

 listed as one quarter the optimum amount 

 during earlier stages and one eighth during 

 the late larval stages. This one fourth, 

 one eighth or whatever it may have been 

 nimierically, was, at any rate, as small an 

 amount of food as was compatible with 

 mere life. Our object was not that any 

 of the larvffi should die of starvation, but 

 that they should live to tell the tale of the 

 results of diminished nourishment. This 

 differeu(3e in feeding was not regulated by 

 lengthening the intervals between meals, _ 

 but by giving the under-fed but a scanty 

 share of the quantity afforded the well-fed 

 individuals at each meal. There were no 



intervals between the meals of the well-fed, 

 whei-eas there were lengthy intervals be- 

 tween the meals of the under-fed, because 

 the under-fed very promptly ate their 

 allotments and were, therefore, without 

 food during the remainder of the interval 

 preceding the next feeding time. 



Tliese experiments have extended over a 

 period of three years, covering as many 

 generations of the insect. The data gath- 

 ered (being the measurements, weight and 

 duration of each larva in each of its five 

 stages ; the time of spinning, weight of silk 

 and weight and duration of each pupa; 

 and the weight, size, pattern and fertility 

 of female of each imago) furnish material, 

 then, for a study of the effects of under- 

 feeding upon individuals during a single 

 generation (the 1903 generation or that of 

 1902 or 1901), during two successive gener- 

 ations (1901-02 or 1902-03), and two alter- 

 nating generations (1901, 1903) and dur- 

 ing three generations (1901-03), a control 

 lot having been carried for each experi- 

 mental lot so that what is modified may 

 confidently be distinguished from what is 

 normal. 



The practice of isolating the larviE indi- 

 vidually has been observed for some of the 

 lots of 1901 and for all the individuals in 

 all the lots of 1902 and 1903. The necessity 

 for such an arrangement will be appre- 

 ciated by making comparison of a lot of 

 isolated individuals with a lot of individ- 

 uals getting a living in a single tray where 

 competition became -a factor, the amount 

 of food per capita being identical for the 

 two lots. 



In 1901, two lots, each consisting of 

 twenty larvaj, were reared on very short 

 rations, the first lot having its individuals 

 isolated, the second having all of its indi- 

 viduals in a single tray. The amoi^nt of 

 food per capita allowed these two lots was 

 identical— an amount calculated to pro- 

 duce dwarfs. After the second moult, 



