166 Host Plants o/Oscinclla frit among Grasses 



Experiment I. The following grasses having given positive results in 

 the winter of 1919 and spring of 1920 were grouped in one cage: (1) Alope- 

 curus myosuroides, (2) Festuca pratensis, (3) Lolium italicum, (4) Lolium 

 perenne, (5) Poa annua, (6) Arrhenatherum avenaceum. These grasses 

 were infected on 16. viii. 20 with 148 flies, the pots being caged separately 

 on 2. iii. 21. Poa annua and Alopecurus myosuroides, being annuals, had 

 died out by March, 1921. The plants showed signs of attack in the autumn, 

 but the larvae failed to subsist either in the debris or the soil after the 

 death of the host plants. 



Only 12 flies emerged in the spring of 1921 — from Lolium perenne, 

 1 fly after 287 days and from Arrhenatherum. avenaceum 1, 2, 5 and 3 flies 

 after 267, 276, 287 and 295 days respectively, reckoning from the date 

 of infection. 



Experiment II. Holcus lanatus, Bromus sterilis, Dactylus glomeratus, 

 Phleum pratense, Hordeum murinum, Arrhenatherum avenaceum var. 

 bulbosum and Avena flavescens gave negative results in 1919-20, although 

 it has been recorded that 0. frit may oviposit in spring on these grasses. 

 Having been grouped under one cage, these grasses were infected with 

 175 flies in 16. viii. 20 and caged separately on 2. iii. 21. 



Again a very small number of flies, namely 13, emerged in the spring 

 of 1921, but here they were spread over four host plants. Holcus lanatus^ 

 produced 1 fly after 295 days, Bromus sterilis 1 fly after 267 days, 

 Dactylus glomeratus 1 fly after 276 days and 2 flies after 287 days, while 

 from A. avenaceum, var. bulbosum 8 flies were obtained, namely 1, 3, 3 

 and 1 after 267, 276, 287 and 295 days respectively. 



The minimum, maximum and mean periods required for the pro- 

 duction of the spring generation were 267, 295 and 284 days respectively, 

 the average period being 30 days longer in 1920-21 than in 1919-20, 

 although the dates of infection were practically the same in the two 

 cases. 



It is of interest to note that in each of these experiments the majority 

 of the flies emerged from the Arrhenatherum species, approximately 

 90 per cent, being obtained from A. avenaceum in the first experiment and 

 70 per cent, from its variety bulbosum in the second. 



^ Exp. III. Small quantities of the following grasses, grown from seed, were caged 

 separately in August, 1!(20: Agropyron rejiens, Agroj)yron caninnm, Agrostis alba, Aira 

 caeftpitosa, Anthoxanthum odoratum, Alopecurus pratensi.'i, Bromus mollis, Btomus erectus, 

 Cynosurus cristatus, Festuca pratensis, F. sciuroides, F. ovina, Poa pratensis and P. trivialis. 

 Twenty-five flies were introduced into each cage between Aug. 18th to 31st, but in the spring 

 of 1021 the results were negative. A spare pot of Holcus lanatus placed in this series pro- 

 duced one fly on 1.'}. vi. 29. 



