Studies of fluctuations in insect populations VIII The wheat blossom midges on Broadbalk, 1932-40, with a discussion of results obtained 1927-40

This is the continuation of the investigation started in 1927 of the wheat blossom midges, Contarinia tritici Kirby and Sitodiplosis mosellana Gehin, as they occur on the field of permanent wheat (Broadbalk) at Rothamsted Experimental Station. 2. Data are given for the years 1932-40 relating to their biology, viz. emergence, number of larvae per infested grain, larvae overwintering more than one winter and alternative host plants, as well as to the degree of infestation of the wheat and the relative parasitism of the midges. 3. The results of the whole fourteen-year continuous study are considered. 4. The date of emergence of the midges is correlated with the wheat-ear emergence, and it appears that both are correlated with the harvesting dates. Thus knowing the dates of either the midge's emergence or the wheat-ear emergence, one can apparently predict in June the date of harvest. 5. There seem to be cycles (about five years) of abundance of C. tritici and S. mosellana, either considered separately or together. 6. Years in which there are high numbers of wheat midge larvae are followed by years of high relative parasitism and years of low larval number are followed by years of low relative parasitism in both species of midge. 7. High relative parasitism accompanies low winter survival of midges and parasites, low relative parasitism accompanies high winter survival. 8. The percentage grain attack is negatively correlated with the yield of wheat. If there is no adequate compensation in the wheat ear that has been attacked between the time of the attack and harvest, then the midges must be considered as pests. If on the other hand there is compensation, the numbers of midge larvae can only be regarded as an indication of the size of yield or in other words a measure of those climatic conditions during the previous months which influence the yield of wheat. 9. Manuring does not affect the intensity of midge attack. 10. The effect of one year's fallowing is to reduce the infestation by the midges, although this is somewhat masked on Broadbalk field by positional differences in infestation. One part of the field is always comparatively more heavily infested than the rest. This effect of fallowing seems to have disappeared by the third successive crop, i.e. non-rotation increases the infestation by both midges. 11. The effect of one year's fallowing on the number of grains of wheat is a marked increase in the first crop after the fallow. The number of grains in the second crop after fallow, while considerably less, is appreciably greater than that of the third and fourth crops. 12. In the years when the number of blind spikelets of wheat is high, the number of possible grains and the actual number of grains formed are low. Thus by estimating the number of blind spikelets between 5 and 16 July one can obtain an early estimate of the number of grains per ear at harvest. But unfortunately the number of grains formed per ear does not necessarily indicate the yield, since the yield would be measured by the number of ears or grains per acre and the weight per grain.