SEED-PRODUCTION IN VIOLETS 157 



the trees shading it being blown down or felled, humble and hive 

 bees make their appearance in search of honey, and, buzzing from 

 flower to flower, cross one with another. In such circumstances 

 the open sweet-scented violet-blossoms are in request, and the 

 same plant-individual, which for years in the dark shade has 

 developed none but cleistogamous flowers, is now r stimulated by the 

 sun's rays into producing flowers with expanded petals." 



Lord Avebury (British Wild Flowers in Relation to Insects, p. 58), 



after having explained the uses of cleistogamous flowers to the 

 order Violacea, says : — " In fact, according to Bentham, the pansy 

 (F. tricolor) is the only one of our English species in which the 

 showy flowers generally produce seed." With regard to this plant 

 Mr. Darwin writes: — " It is curious in this case, as in many others, 

 how long the flowers may be watched without seeing one visited by 

 an insect. During one summer I repeatedly watched some large 

 clumps of heartsease many times daily for a fortnight before I saw 

 a humble bee at work. During another summer I did the same, 

 and then one day, as well as on two succeeding days, I saw a dark- 

 coloured humble bee visiting almost every flower in several clumps, 

 and after a few days almost all the flowers withered and produced 



fine capsules." 



In the spring and summer of 1905 I originated a series of ex- 

 periments, hoping thereby to ascertain whether the early open 

 flowers of our British violets are useful as well as ornamental. A 

 wild-flower garden at Fairleigh, Weston-super-Mare, where I culti- 

 vated almost all our violets (excepting those which need marsh or 

 fen) greatly facilitated my efforts. A sunny bank, facing south, 

 helped the bees to do their duty, although 1 must admit that they 

 were more often occupied in visiting plants of the Boiaginacea. 



As the showy open flowers unfolded, I tied red silk round the 

 stems of V. odorata vars. sub-camea and alba, a red violet (which I 

 suspect of being a cross between F. sub-carn$a and ordinary V. hirta), 

 F. hirta, F. hirta var. propera (of which the seedlings were notice- 

 ablyliairy), T. niviniana, and F. silrestris. The two latter were 

 less disposed to sSt seed than were the acaulescent violets, though 

 the seed set by them germinated more quickly. On May 4th I tied 

 muslin bags around the now perfectly formed capsules, and later 

 carefully gathered in my seed, which was placed at once in paper- 

 bags and labelled. 



In July of the same year, by the kindness of Mr. Lynch (the 

 curator), I was allowed to sow the seeds in boxes protected by a 

 frame in the Botanic Gardens, Cambridge. Except in the cases of 

 two species, I counted in the seeds, by which means it was possible 

 to arrive at pretty accurate results, which were as follows : 



Seeds Sown. Seedlings. 



1. F. odorata, sub-camea... 12 12 



2. F. odorata, alba 12 8 



3. V. odorata (red) 30 3 



4. V. hirta 1 1 



5. V.Jdrta, propera Many Many 



6. F. Hiviniana Several 12 



