208 fn/ierifanre of GJiiiue Lemjtli and of Colour iu Oats 



occurring when the grain is at its plumpest, appears hirgely to deter- 

 mine the size and intensity of the streak. No doubt this also accounts 

 for the fact that only some of the grains in the ears of a streaked plant 

 actually bear the streak, as but few get exposed in this way, in the 

 long-glumed plants particularly. That direct sunlight exposure does 

 not alone determine colour production, however, is proved by the 

 existence of the flushed forms, wherein the greatest colour area is de- 

 veloped in well covered np portions of tlie grain. The action of direct 

 light may start in these grains the chemical change which renders the 

 colour visible, but that it is essential for its continuance is a matter of 

 doTibt. 



. On the basis of the colour classification already indicated the F^ 

 generation consisted of 



28 Flushed, 8 Streaked, 136 Non-coloured. 



Or, eliminating all which gave less than F-, plants next season, 



27 Flushed, 8 Streaked, 123 Non-coloured. 



The nature of these Fa plants was determined firstly by examination 

 of their grain in the ordinary way, and subsequently, after the more 

 accurate acid test had been adopted, by re-examination of any seed 

 left over from sowings for the F^ rows, or, when none remained, by 

 considering the i^, segregations analogically. 



Thus, in the case of the 8 streakeds, only one (which had a very 

 strong streak) was detected on rubbing out by the rough, hurried, 

 unaided eye test then used. Later, after the sowings for the F.. 

 generation had been made, the grain left over from three of these was 

 tested in acid and streaks were discovered microscopically. Now as no 

 grain remained over from the other four F-,, sowings, that the streak was 

 present in their F. parents was deduced from the fact that the i^rogeny 

 showed exactly the .same type of segregation as that of the undoubted 

 streaked forms. 



In the same way the 123 non-coloureds were definitely identified as 

 such. After sowing, surplus .seed was left over from 37 F^ plants. It 

 averaged about 8 grains per plant, and, coming from short-, inter- 

 mediate- and long-glumed forms indiscriminately, constituted a good 

 random sample. All these grains gave a negative colour reaction in 

 sulphuric acid, and the analogous segregation in the F^ rows of the 

 remaining 86, — roughly eye-tested even though no grains were left 

 over for subsequent microtesting, — settled the correct classification of 

 the whole groiip. 



