PRIMROSES AND THEIR CULTIVATION'. 



379 



— rich imperial purple — is good. About four years ago I had a 

 grand specimen of this plant on a raised bed in ordinary garden 

 soil ; it bore two umbels of thirty or forty flowers each in April, 

 and again in September. A portrait of the autumn flowering 

 was taken by Miss North and engraved for the Garden. Though 

 I have done my best to imitate the conditions, I have never pro- 

 duced anything like this again. 



As for other Himalayan Primroses, I can add but little to what 

 I said four years ago, as recorded on page 260 of the Report of 

 the Primula Conference. It is desirable to retard the flowering of 

 P. dcnticulata and its varieties, in order that it may be postponed 

 until the conditions of our English climate give it a better chance 

 than it has if it tries to flower at Christmas. I have discovered 

 no better plan than cutting up the large crowns in August, as 

 before recommended. By exercising selection, having regard 

 both to colour and to habit, the varieties of this species are 

 capable of improvement. If by crossing we could improve the 

 colour and size of the flower, and retain the great vigour of con- 

 stitution, this plant might have a future before it — a cross I 

 made with P. rosea was fit only for the rubbish heap. As an in- 

 stance of the prolific habit of the species, I may mention that 

 one which accidentally lost its head about two inches below the 

 surface produced the following spring about twenty shoots from 

 the neck of each main root. These appeared like a ring of seed- 

 lings, about 200 in number, each attached to the old roots by a 

 thin rootstock like thread ; some of them which I tried for ex- 

 periment made good flowering plants by autumn. This mode of 

 multiplication may be utilised by studying the required conditions. 

 Closely allied species or varieties are in cultivation under the 

 names P. Fortunci, P. cashmeriana. P. erosa, Sec. 



P. rosea has a thickly tangled mass of fibrous roots, showing 

 that it requires a finely pulverised soil. The easiest way of deal- 

 ing with it is to make colonies during summer — any time 

 between the ripening of the seed and the end of August — by 

 pulling a plant to pieces, breaking off all the old root, and 

 planting the shoots singly wherever the soil is suitable. They 

 must, of course, be frequently watered till established, and no 

 colony must be expected to do well for more than two years — 

 renewal every year is better. But seedlings from well- selected 

 seed are more satisfactory still. No successful hybrid has yet 



l2 



