38 



ILLINOIS BIOLOGICAL MONOGRAPHS [280 



a few other members of this family (Liparidae). This peculiar condition 

 has been observed by the author in but one nocluid species, Caenurgia 

 ereclilea. Its significance will be discussed later. 



Besides these environmental and sexual variations in the number of 

 stadia, there are very probably hereditary tendencies toward individual 

 differences in this respect. Davis records one individual of Cirphis \ini- 

 puncta passing seven stadia instead of the usual six. Since this exceptional 

 individual was reared under the same conditions as hundreds of others, it 

 seems evident that heredity and not environment must account for this 

 exceptional instance, the possible significance of which will be considered 

 later. Similarly, one larva of Agrotis ypsilon, reared with fifty-two others, 

 molted but six times instead of seven, according to our records. 



In 1890 Dyar called attention to the fact that the widths of the success- 

 ive heads of any lepiJopterous larva in all its stages bore a certain definite 

 relation to each other. His presentation of this point may be summarized 

 as follows: the quotient obtained by dividing the width of the head of any 

 instar by that of the previous one is a constant, which is characteristic 

 for the species. This principle has been termed Dyar's Law. Its utilitar- 

 ian value is obvious, enabling one to determine what instar he is dealing 

 with when a specimen or a published measurement of any other known in- 

 star is available. A fair indication as to the number of stadia may also be 

 obtained if the size of the first and of the last instar is known. 



An inspection of a large number of species will reveal the condition in 

 this regard within the Noctuidae. Although the measurements represent 

 averages derived from the number of individuals indicated, in many cases 

 a much larger number has been examined to insure the determination of 

 a fair average, as well as to find the extremes of variation. The material 

 studied was either preserved when collected or grown under approximately 

 natural conditions. Individual variation in size is not as great as might 

 perhaps be expected, usually rendering the identification of instars a 

 simple matter. The figures expressing the percentage of variation are ob- 

 tained by dividing the maximum variation found, by the average, multi- 

 plying by 100, (to express as percentage) and dividing by 2, so that the 

 deviation from the mean in either direction, not in total, is represented. 

 The later instars naturally present the greatest variation, having been 

 longer subject to external influences. Where the measurements are based 

 on individuals of different broods the variation is usually larger than other- 

 wise, since the larvae of certain generations often grow larger than those of 

 others in an ordinary season. All measurements have been made with an 

 ocular micrometer. 



The inconstancy of Dyar's supposed constant, which we will refer to 

 as the index of growth, is striking, varying in Agrotis ypsilon,ior instance, 

 from 1.28 to 1.84. The average for any particular species ranges from 1.44 



