OBSERVATIONS ON INDIAN PRIMULAS. 



215 



In point of colour of flower there is less variability than in shape of 

 flower, form of leaf, and degree of mealiness. The alpine states have as 

 a rule much larger flowers than those of lower altitudes. And what is 

 more surprising still, while the low-level forms are seen to originate 

 large clumps, within crevices of rocks in damp situations or even under 

 the spray of waterfalls, the alpine forms prefer the shade of bamboo or 

 pine. They invariably form large clumps, often a foot or more in 

 diameter, and are seen very frequently as one mass of bright rose-purple 

 to pale lilac flowers with yellow throats. But there is still another 

 circumstance that I think it is well to mention. The whole of the 

 clumps in one neighbourhood flower simultaneously and have repeated 

 flushings throughout the year. On April 20 I passed along the 

 Toungloo range in Sikkim when P. pctiolaris was a blaze of flowers. 

 I returned twenty days later along the same path and could not 

 discover a single flower. I have collected it in flower from March 

 to September, though it is best in May and June. 



I have gone into these details with P. pctiolaris because I think 

 it is a much-neglected beautiful species. It sports almost too freely, 

 is perennial, and easily grown if liberally supplied with water or 

 planted alongside of limestone rocks. I have another reason : it 

 is very largely representative of the series with which I am at present 

 dealing. To understand them fully, however, it is desirable that I 

 should classify the petiolaris section a little more in detail. There may 

 be said to be two great subsections : — 



1. Leaves glabrous or nearly so, and ovate-spathulate to subrotund, 

 sheathing. The species are P. petiolaris in all its forms, also P. 

 Stirtoniana and P. Hookeri. Then a special group that have large 

 sheaths on an erect stem, viz. P. pulchra and P. Dyeriana. Lastly two 

 plants, P. Clarkei and P. reptans, that I place here because of their 

 being glabrous N.W. Himalayan forms, but they might otherwise 

 more naturally be assigned positions in the set that follows. 



2. Leaves puberulous or tomentose, rotund, and possessed of 

 distinct petioles. The examples are P. reticulata, P. rotundifolia, 

 P. Gambleiana, P. Forbesii, P. Listcri, P. vaginata, P. mollis, and 

 P. geranii folia. 



I desire to .bring before you only the more striking facts, of the 

 classification with a view to establishing general principles that I think 

 may be of practical value. No single character in the above diagnostic 

 separation would by itself be of any real value, but when taken in the 

 aggregate they are of great assistance. For example, were we to 

 simplify the above into glabrous and puberulous forms we should 

 instantly meet with numerous stumbling-blocks. I need but mention 

 that P. reticulata, P. rotundifolia, and P. Gambleiana are often almost 

 glabrous. So again leaves spathulate and leaves petiolate would be 

 useless characters by themselves, since, as I have just said, both 

 conditions may be met with on one and the same plant. Still again, 

 flowers solitary or flowers umbellate would be quite misleading, since 

 both these conditions may be seen on the same root. But when we 



