INFLUENCE "I -IARVATIOX OF INSECTS OX OFFSPRING. 23 



< aterpillars were mated to control moths which were at my dis- 

 posal. In the filial generation the conditions of rearing were 

 identical in each lot. Food was given daily and in superfluity. 

 A- .ill lot- originated from one pair of moths caught in 1917. 

 the material is to a certain degree genetically homogeneous. 



Tin following physiological characters of the filial generation 

 uere taken into account: (l) The duration of the larval life, 

 \\hich in the parental generation had undergone considerable 

 prolongation owing to inanition, (2) the duration of the pupal 

 stage which in the parents had undergone a certain abbreviation, 

 (3) the limits of growth of the caterpillars expressed by the weight 

 of fresh pupa- which in the starved parental generation were 

 much smaller than control specimens. Apart from this, the sixe 

 of the eggs out of which the first filial generation developed, 

 ilicir number and their capacity to develop as well as the mortal- 

 ity of caterpillars and of chrysalids have been examined. The 

 changes of the above-mentioned characters were in the filial 

 generation in part much smaller than in tin- parent-. I therefore 

 did not use for my calculations in these experiments the averages 

 of a character for each lot separately, but in each experiment I 

 joined .ill specimens of either sex of all lots into class-frequencies 

 as usual and I calculated, the average values for all specimens 

 together with the probable error of these averages (A E,\). On 

 comparing .my character of the control and of the experimental 

 material only such differences are considered as essential, i.e., 

 biometrically well grounded, which are larger than four times 



... / DifT. \ Di IT. 



their probable error I = > 4 \ When the ratio is < 4 



V Enut. i tL.\i\n. 



the differences are considered as biometrically not essential and 

 the character as not changed. 



Tin-: OFFSPRING OF FEMALES Sunji-: IM \^ si \m \TION M.\n:i 



\\ 1 1 ii ("<>\ i k< ii M M.I-IS. 



From Table I. we may infer that the number of eggs laid by 

 these specimens was much smaller than that prodmed by control 

 females mated with control male-. < )n the other hand, the sixe 

 of the eggs was biometrically identir.d in both cases, as is evi- 

 denced in Table II. Almost .ill eggs from -t.ir\ed females were 

 i .1 |).ible of development, although a large number of the develop- 



