THE SORGHUM MIDGE. 55 



the winter it is no uncommon thing to find them within the seed 

 glumes, but the writer has not found a single instance in which nor- 

 mal larvse occurred in heads that had been formed during the summer 

 and allowed to stand for a sufficiently long time to yield midges from 

 all possible infestation; only occasional naked pupae and a predomi- 

 nance of "cocooned larva?" are contained in the latter. 



These "cocooned larvae" are formed in the seed during the entire 

 breeding season of the midge. As early as June, in the latitude of 

 San Antonio, Tex., these forms were found, and there seems to be no 

 regularity in their habit of emerging. After all emergence has ceased 

 from normal naked larva? these forms develop into cocooned pupa? 

 and emerge irregularly. Ofttimes a cocooned pupa is found within 

 a seed along with a normal naked larva. Just what controls the 

 development of this form has not been discovered. Attempts to 

 induce hibernation artificially by subjecting these to low temperatures 

 and later placing them in a warm room have been unsuccessful. 



RELATION OF JOHNSON GRASS TO THE MIDGE PROBLEM. 



From what has been said previously in regard to the midge in rela- 

 tion to Johnson grass, it is a self-evident fact that this grass furnishes 

 the key to the situation. Johnson grass allowed to remain over win- 

 ter in and about sorghum fields carries the midge until spring, and 

 being the first to head and bloom, gives the midge a good start, and 

 by the time the sorghum is headed there is a large brood of midges 

 from the grass ready to infest it. (See PI. I, fig. 1.) Johnson grass 

 is generally considered one of the greatest pests on the farm, and i|s 

 function as a host for the sorghum midge serves as but another indict- 

 ment against it. It is no uncommon sight to find sorghum fields 

 from which the last crop has been harvested, with Johnson grass 

 growing and heading in the fence corners (PL II, fig. 3) and even in 

 the fields (PL II, fig. 2). 



NATURAL ENEMIES. 

 PARASITES. 



The midge in certain localities is abundantly parasitized by a 

 small black hymenopterous parasite determined as Aprostocetus 

 diplosidis Crawf. (figs. 29, 30) and by a smaller parasite determined 

 by Mr. Crawford as Tetrastichus sp. (fig. 31). The latter is known 

 to be both primary and secondary, but it is more likely to be primary 

 in its relation to the midge. These latter parasites are reared from 

 infested sorghum heads along with the predominant Aprostocetus 

 diplosidis. Although the predominant parasite is very aggressive 

 and parasitizes the midge very actively it does not become suffi- 

 ciently numerous to materially check the midge until late in the 



