An account is given of the biology and importance of sorghum midge, Contarinia sorghicola (Coq.), in Ghana. It is distributed throughout the interior savannah areas and because of traffic in grain sorghum (guinea corn) has spread into forest localities and some places in the coastal savannahs. Host-plants comprise all varieties of Sorghum vulgare sens. lat. and the grass Sorghum arundinaceum. Its occurrence in Pennisetum polystachyon has not been confirmed.

The biology of C. sorghicola in Ghana does not differ significantly from that described elsewhere. Adult emergence occurs early in the morning and peak oviposition is from 0800 to 1000 hr., shortly after emergence. The majority of spikelets contain only one larva or pupa, but 64 per cent. of immature stages (larvae + pupae) nevertheless occur two or more per spikelet. Density per spikelet is related to rate of infestation, with maximum density at about 2·5 immature stages per spikelet at infestation rates of about 40 per cent. The mean density over two seasons was l·7 immature stages per spikelet. The cycle from egg to adult requires 17 to 21 days to complete.

The onset of diapause in larvae and the eventual breaking of diapause may be the consequences of threshold reactions to moisture and temperature, with a cumulative process operating in breaking diapause. It is suggested that observed differences between northern and southern localities in dates of appearance of adults and subsequent infestations indicate that moisture is the critical threshold factor in the north and temperature in the south.

Parasites are of little importance. Only Eupelmus popa Gir. has been identified from Ghana, and it exerts no influence on population levels.

A most important relation exists between midge infestation and time of main flowering of the crop. When main flowering is early in the season, midge attack is very low, below 5 per cent. or even non-existent, but as dates of flowering are delayed so the percentage of spikelets infested rises rapidly, an infestation of 81 per cent, having been recorded in a late crop. In general, it seems that infestation is heavier in the less severe climate of the Guinea savannah than in the Sudan savannah areas.

A sampling method is described in which three categories, ‘sound’, ‘midge attacked’ and ‘others’ (which included all undersized or damaged grains not resulting from midge attack) were set up. In respect of midge, it was shown that attack was evenly distributed over the head of sorghum but that in some cases the ‘others’ category gave highest rates at the bottom of the head. This, it is suggested, may be due to differential attack by various species of Hemiptera, which may be responsible for a considerable proportion of the grains included in the ‘others’ class. Surveys carried out in 1952 and 1953 showed that midge was not a very serious problem, infestation rates above 10–15 per cent, being unusual except in late-flowering crops. On the other hand, losses hi the ‘others’ category were rarely below 15 per cent. and were often well over 30 per cent., indicating a major source of loss.

It is confirmed that Nunaba, which typifies the group of varieties in which the glumes remain closed at anthesis, is resistant to midge. It is also confirmed that this resistance breaks down in the absence of a more favoured alternative variety. There is evidence that susceptibility to loss in the ‘others’ category varies between some Nunaba × Belko crosses, one cross, AA226, showing an appreciably smaller loss than two others; the variety Kamolgu may also be less susceptible.

It is suggested oft the basis of field observation in Ghana and references to literature that Hemiptera are the primary cause of loss in the ‘others’ class; in Ghana, Riptortus spp. and Mirperus spp. are important. If this damage is consistently serious, the problem of measures needed to reduce losses is complicated. Midge could, in the long run, be controlled or even eliminated as an economic pest by adoption of practices leading to early flowering in periods before the appearance of midge. In Ghana, these periods are given for the main sorghum areas of the north and north-east, and it is suggested that development of varieties of shorter maturity periods would assist by easing the congestion at normal sowing times in May and June. There is no significant relation between percentages recorded in the ‘others’ category and flowering date, so that other methods of control must be looked for, and it is suggested that search for resistant, or less susceptible, varieties offers a promising avenue of approach. Evidence is presented that losses from both midge and ‘others’ can be reduced by a 1 per cent. DDT spray, but chemical control is unlikely to be of general use.

The most urgent need is for a general investigation of causes of loss in sorghum, with particular attention to the agents other than midge that are responsible for what appear to be large annual losses. Standard trials covering a range of conditions and flowering dates, and aimed at establishing the economic status of the sorghum midge when related to dates of flowering, are proposed. These would also provide much of the information necessary for the formulation of further programmes for the study of the other agents.