120 Observations on the Habits o/Oscinella frit 



been realised, but very little experimental breeding has been recorded 

 in proof thereof and no attempt has been made to estimate the relative 

 preference for different grasses at different periods of the year. 



Method. 



Breeding the flics from infected grasses brought in from the field 

 gave uncertain results, since it was almost impossible to obtain pure 

 clumps, except in very few cases, e.g. Lolium spp. and Alopecurus 

 myosuroides. Sowing seed in small patches in the field was also ineffective. 

 Pot experiments gave positive results (detailed in Synopsis II) with some 

 host plants but without regularity. Negative results, therefore, had very 

 little value. 



Samples of common wild and pasture grass seeds were collected and 

 sown separately in 9-inch pots, netted and kept for convenience of 

 observation, in an unheated greenhouse^, with the lights always fully 

 open. 



Flies were reared from field plants, sexed under the binocular micro- 

 scope, and introduced into the cages. The flies on emergence were easy 

 to observe, as on disturbance they congregated at the tops of the cages. 



Two series of plants were infected, one in July and the other in 

 August, 19192. 



There is no evidence that the flies reproduce themselves without 

 previously securing food supplies, and the longevity experiments (p. 123) 

 indicate that food would probably be sought^. If food is essential, the 

 adults in these experiments obtained sufficient from these particular 

 grasses to enable them to reproduce, but probably insufficient to enable 

 them to live a normal period. In 1919, in the cage of L. italicum, the last 

 fly died on October 14th and in the cage of L. perenne, on October 11th 

 {vide Table I for field record). 



The parent flies used in these experiments were not more than 5 days 

 old and the majority less. 



^ An insectary was not available at this date. 



^ Negative results were obtained with the following grasses in both series of experiments : 



Alopecurus pratensis, Arrhenatherum avenaceum var. bulbosum, Avena flavescens, Bromus 

 sterilis, Holcus lanatus, Hordeum murinum, Phleum pratense, Poa pratensis. Six other 

 grasses also gave negative results, but here growth was not considered normal. Further 

 experiments are being conducted this season with all these host plants in large breeding 

 cages under more natural conditions. 



^ On April 4th, 1920, many flies were observed frequenting apple blossom on trees near 

 the oat crop, and in May and June, on the flowers of Veronica hederifolia in the field, 

 apparently feeding. Knuth (W) records O. frit as visiting PotentiUa sylvatica, Daucas 

 carota, Matricaria inodora and Mentha aqitatica. 



