THE TSETSE PROBLEM IN NORTH MOSSURISE. 385 



will reach a considerably higher proportion of pupae than those of the earlier fires, 

 the anticipated effect on the flies on the wing would be well worth obtaining also. 

 Birds, again, as I have shown already, probably attack tsetses most when these are 

 in numbers together, and the temporary clearing of the game from a large area will 

 remove the agency that keeps the smaller tsetses scattered, while hunger and the 

 black background should increase the flies activity and visibility. 



The one thing that has made me hesitate to reconmiend simultaneous burning 

 is a slight doubt as to its exact effect on the insectivorous bird population itself. 

 October burning in any case encroaches somewhat on their breeding season, but not 

 to an important extent — that is, the early nesters that have had their nests destroyed 

 have still plenty of time before them, and even in ordinary years an amazingly large 

 proportion of them has to make a fresh start as the result of the depredations of 

 enemies that are likely to be adversely affected by late burning. Again, a fire and 

 the few days following it are feast days, for a very large proportion of the winged 

 insect population escapes it and, being then more conspicuous against the universal 

 blackening, is more readily detected by the birds, which may sometimes be seen 

 searching the burnt ground in parties. The real point is whether the wholesale 

 destruction everywhere of the insects' own food will so much reduce their numbers 

 as to react adversely on the birds. I think that there will be no difficulty here at 

 all, for in this district at that time of the year the response of the vegetation after 

 burning is immediate and most rapid, and a burn gets repopulated with wonderful 

 speed, the new brood of grasshoppers even showing colour adaption to the scorched 

 surroundings ; but the effect of simultaneous burning on the insectivorous birds 

 should be noted nevertheless. 



In short, the effects of simultaneous burning, while very promising, have yet to be 

 accurately observed ; but the good effects of late burning have actually been demon- 

 strated. 



I wish, therefore, to make it the one outstanding recommendation of this report 

 that late burning — with, very frequently indeed, a year of no burning — should be 

 given a trial over a considerable number of years ; also that, in the first season at any 

 rate, an effort should be made to have the burning as simultaneous as possible over 

 a large block. The rubber forests and finer Lusitu wooding generally could, if 

 desired, be left out of the scheme. Early burning thereabouts, every year, will 

 save them. It will also unfortunately spread both them and the fly. The fine 

 timber there must be considered also ; and against the reduction of the cost of 

 clearing which late fires might have effected by the time this section comes to be 

 settled, must to some extent be placed their destruction of the surface humus. 

 Elsewhere this is being destroyed in any case by the useless fires ; it may as well 

 therefore be destroyed by useful ones. The very deciduous areas at the higher 

 elevations need not necessarily be included either, but they are not sufficiently 

 extensive to be worth excluding, especially as their wooding (for the most part of little 

 value) will remain a great annual danger so long as any permanent fly survives below. 



To give the measure a definite experimental value sample plots or areas in each 

 type of wooding should be selected, described, charted and photographed. These, 

 examined from time to time, would afford a measure of the success or failure of the 

 experiment. 



