A study of a natural population of Phytodecta olivacea (Forster) (Coleoptera, Chrysomeloidea)

Abstract

This is a study of the seasonal and annual changes in a self-contained natural population of the Chrysomelid beetle, Phytodecta olivacea (Forster) which lives on broom, Sarothamnus scoparius (L.) Wimm. The adult beetles emerge in the spring from hibernation in the soil and after laying their eggs on the food-plant re-enter the soil where between one-third and one-fifth of them survive to the next summer. Their offspring on the broom become adult in the late summer and after a short period of feeding join their parents in hibernation. The adults rarely fly and dispersal seems to take place mainly by walking. Regular estimates of the numbers of the insect were obtained by sampling the food-plant, the quantity of which was known in terms of standardized ‘armfuls’ and which could be related to weighed armfuls. Adult beetles were estimated by shaking armfuls over a tray and eggs and larvae by examining sprigs of broom of known weight. Some records of the numbers of adult beetles are available from 1948-59, but detailed estimates of all stages were kept only in 1954-58. The numbers of adult beetles in 100 armfuls tended to fall into a negative binomial series, indicating some tendency to aggregation. The total number of adults which emerged in the spring is equivalent to the total sum of all the weekly samples divided by the effective adult life (i.e. the time spent by the beetles on broom) which was determined in the laboratory. This total emergence was independently estimated by counting the number of beetles which were found in forty special enclosures in the field. The number of eggs laid each week could be estimated by knowing the number of females present (from counts in armfuls) and by the use of a regression equation, calculated each year in the laboratory, relating the oviposition rate to the age of the population and the mean temperature. The total number of eggs laid in the season was the sum of the weekly figures. Assuming that the destruction of eggs was exponential, it was possible to deduce the mortality from the initial numbers (based on the regression equations) and the duration of the stage at the prevalent temperature. The relation of egg and larval stages to temperature had earlier been established experimentally (Waloff & Richards 1958). The mortality in the larvae was similarly deduced, treating the surviving eggs as the initial number of first-stage larvae. Adult beetles were destroyed by a braconid parasite, Perilitus dubius (Wesm.), and by a fungus Beauveria bassiana (Bals.) Vuill. Destruction by the braconid was 5 to 13 % of the spring and 5 to 26 % of the autumn beetles. The fungus destroyed the beetle chiefly during hibernation. In dissections, 0 to 25 % of the beetles were found to be infected in different years, but there are some grounds for thinking that much of the winter mortality was also due to this cause. Of the eggs laid in the field, 4 to 10 % were found to be sterile and less than 1 % were parasitized by Hymenoptera. Between 1 and 3 % of the larvae were parasitized by the fly, Meigenia mutabilis (Fall.), and another small fraction by various other parasites. Most of the mortality of the eggs and larvae was due to predatory insects, especially Hemiptera Heteroptera (Miridae, Anthocoridae, Nabidae) and earwigs (Dermaptera). The losses due to these causes were confirmed quantitatively in 1958 and 1959 by using a serological method to detect the food of the predators (Dempster 1960). There was also an additional small mortality of the pupae in the soil due to carabid beetles and possibly other predators. The effects of most of these causes of mortality were made more striking by concentrations of the populations which occurred in 1956 and 1958 after about half of all the broom had died in the preceding winters. Many of the plants had reached the end of their natural life (10 to 15 years) and had become susceptible to frost and unable to survive a heavy production of pods. On the reduced quantity of broom, the number of beetles per armful in 1956 was about twice as great as in the year before while the numbers of alternative prey, such as aphids and psyllids, were lower than in subsequent years. The calculated mortality of the eggs and larvae in 1956 was 99 % compared with 79 % in 1955. Similarly, the mortality of the early stages was 92 % in 1958 compared with 78% in 1957. Although one or other species of predator is found throughout the season, the most abundant, the Miridae, hatch early and their numbers decline sharply at about the time their adults appear. Thus, Phytodecta suffers from much more predation in the early part of the season than later, after the middle of June. In 1955, eggs had been found in the field for 6 weeks before any larvae were found. In general, the early stages of Phytodecta whatever other disadvantages they may suffer have a better chance of avoiding predation in the second half of the summer. Temperature and rainfall are the two climatic factors which most influence Phytodecta . Temperature has an important influence on the oviposition rate and the duration of the immature stages and seems also to alter the length of time spent above ground by the adults. It also has indirect effects through parasitism and predation. Rainfall during the period of larval development influences the size of adult population in the following year. These observations on the changes in the population are summed up in life-tables for the years 1954-58. In short-lived species these would be best described as ‘budgets’, since they deal directly with the causes limiting the growth of the population and not with the age distribution within it. Each cause of mortality is expressed as a percentage of the stage on which it acts and as a percentage of the total mortality in the complete annual cycle. The second method of expression gives a better measure of the importance of any factor because a large apparent mortality late in the cycle may be only a small fraction of the initial numbers. By using this method, it is found that about 78 to 99 % of all mortality occurs in the eggs and larvae on broom; 0.3 to nearly 19 % occurs in the later stages in the soil in the early autumn, and 0.4 to 4.5% in the adults in the soil in the winter and spring. In the adult stage, deaths due to hymenopterous and fungal parasites if added to the numbers of beetles known to survive until the following year, leave a certain proportion unaccounted for (5 to 42 % according to the year). This discrepancy is probably largely due to additional deaths from fungal attack during the winter. In the eggs and larvae, while disappearance from the broom was recorded each year, only in 1957-58 was predation, the principal cause, measured; in these 2 years it apparently accounted for 71 and 88 %, respectively, of the disappearance. In these stages, some of this discrepancy is due to parasitism and to sterility of the eggs. Although all these estimates are subject to large sampling errors, we find some support for them when we compare the difference between actual mortality and that necessary for a stable population with the actual trends in the population from year to year. In four years the trend was in the right direction and only in 1957-58 did an increase of population follow an estimated excess mortality. The life-cycle of Phytodecta olivacea has one feature which makes it partially independent of the mortality in any one year. A variable proportion of the adults survive to reproduce a second time and a high mortality or ineffective reproduction during one summer can be partly compensated if a high proportion of the parents survive. Further, both emergence of the adults in the spring and reproduction are relatively protracted, so that each stage has a chance of avoiding some of the seasonal hazards, such as predation or unsuitable weather.