Contrasting dynamics from egg to adult in the life cycle of summer and overwintering generations of Baetis rhodani in a small stream

Summary 1. The chief objectives were as follows: (i) to describe and compare quantitatively the life cycles of the summer and overwintering generations of Baetis rhodani , including laboratory experiments to determine the number of eggs laid per female and the number of larval instars; (ii) to test Dyar’s hypothesis; and (iii) to test for density dependence in the life cycle. Samples of larvae and downstream‐drifting adults (newly emerged subimagines and spent female imagines) and counts of egg masses were taken every 2 weeks over 39 months in a small stony stream. 2. Although females from the summer generation were smaller than those from the overwintering generation and laid fewer eggs, the relationship between the number of eggs laid by a female and body length was the same for both generations, being described by a power function. An exponential equation described the relationship between the mean body length of each instar and instar number; hence, Dyar’s hypothesis was validated. Adult males emerged after 17 or 18 larval instars and females after 18 or 19 instars in the summer generation, with higher values of 25 or 26 instars for males and 26 or 27 instars for females in the overwintering generation. Mean instar body length in the summer generation was always larger than that in the overwintering generation, but this did not produce larger final instar larvae because the fewer instars resulted in males and females being smaller on emergence than in the overwintering generation. 3. Male subimagines started to emerge before females, and their emergence patterns indicated that there were two cohorts in each generation. Ovipositing females in the first cohort were always larger than those in the second cohort. Therefore, four mean values for eggs/female (=eggs/egg mass) were used to convert egg masses per m 2 to eggs laid per m 2 . Larvae were divided into six stages: I = instar 1, II = instars 2–5, III = instars 6–10, IV = instars 11–16, V = instars 17–24 (overwintering generation only) and VI = final instars prior to emergence. There were two clear modal densities for each larval stage in each generation, confirming the presence of two cohorts. 4. These modal densities were used to estimate loss rates between successive larval stages, and ‘ k ‐factor analysis’ was used to examine the relationship between loss rates for a particular larval stage and the modal value for eggs or larvae at the start of that stage. Loss rates in larval stage II were density dependent and were when population regulation occurred. For all other larval stages, loss rates were density independent and fairly constant with no significant differences between the two generations. Life tables were constructed for egg densities close to the range of modal values found in the stream (1332–11 512 m −2 ). The egg density that yielded the maximum number of final instar larvae, and hence adults, was 2700–2750 m −2 with poorer survival for initial egg densities below and above these values. 5. This study has shown that (i) the mechanism for population regulation was density‐dependent survival in larval stage II (instars 2–5); (ii) the importance of obtaining information on egg densities and their role in density‐dependent processes; (iii) the necessity of frequent sampling for a quantitative description of complex life cycles of aquatic insects.