In the study on population dynamics of mosquito larvae, it is necessary to know about age structure of larval population as well as population size. The purpose of this paper is to elucidate which of the following six morphological characters-head width, body length, siphon length, number of labial teeth, of lateral scales and of pecten teeth-is suitable for determination of larval ages (instars) of mosquitoes. The results obtained are as follows : 1) As to the larvae reared individually, the most suitable characters for determination of instars are both head width and siphon length in Culex pipiens s. 1. and head width in Aedes albopictus. The conclusion was made in the following way : Observation was made on each datum of the six characters mentioned above and four frequency distributions were found. These distributions, representing the normal patterns, corresponded to four larval stages; from the first to the fourth instars, respectively (Fig. 2). Then, suitability of these characters for determination of larval instars was tested quantitatively by means of the "instar determination index" as defined below : D(r-1)・r=(the lower rejection limit (α=0.05) of frequency distribution for a certain character of the rth instar)/(the upper rejection limit for the same character of the (r-1)th instar) When the index for a certain character is over 1 (namely, two adjacent distributions of the character are separated from each other), (r-1)th and rth instars can be distinguished; and the larger the index, the more reliable the determination by this character (Table 2). 2) In C. pipiens s. 1. and A. albopictus, four frequency distributions were distinguished in measurements of larval head width of outdoor populations as well as of laboratory colonies reared individually. When arranged in the order of size as shown in figure 3, each of the four distribution ranges (value of mean±rejection limits at 5% level) of the outdoor population agreed well with the corresponding range of the laboratory colony, in which the four larval instars coincided exactly with the four frequency distributions (Fig. 3 and Tables 2, 3). This result suggests that the four frequency distributions of larval head width in the outdoor population obviously correspond to the four larval instars, and that instars of certain larvae of these species can be distinguished by comparing their values of head width with the preliminarily measured ranges of the outdoor larval populations of the same species. 3) Four clear-cut frequency distributions were observed in measurements of larval head width in outdoor populations of the following eleven species : Anopheles hyrcanus sinensis, Culex tritaeniorhynchus, Culex vishnui, Culex orientalis, Culex pipiens s.l., Culex pipiens molestus, Culex vorax, Aedes albopictus. Aedes togoi, Aedes japonicus and Tripteroides bambusa. And the instar determination indices of larval head width calculated as written above were over 1 with no exception (Fig. 4 and Table 3). This suggests that head width is a suitable character for determination of instars of mosquito larvae of at least eleven species belonging to four genera. In addition, this seems to be equally true of Armigeres subalbatus and Uranotaenia bimaculata. The early-instar larvae of these species, however, were not collected during the present investigation. 4) When the logarithmic values of larval head width (y) were plotted against instar (x) in all the species mentioned above, they fell on straight lines, to which the following general formula gave a good fit : log y=a+bx (a, b : constants) (Fig. 5 and Table 4). This fact suggests that larval head width of most species of mosquitoes increases exponentially after each ecdysis and remains with little growth during the stadium. The regularity in growth of larval head width in all the species inspected gives an additional evidence that larval head width is most suitable for

(View PDF for the rest of the abstract.)