256 DJt J. M. MACFARLANE ON THE 



periods, and to search available literature on the subject. This was confined to Mr 

 Lindsay's record* of 1408 plants flowered on the Rock Garden, to a few others given 

 in the same publication over a period of twenty years or more, and largely to unpublished 

 lists of plants flowered on the Rock Garden during the last ten years. During two 

 seasons also (1890 and 1891) several wild hybrids and parent forms have been watched. 



Naturally, for exact comparison of the flowering period of hybrids and of their parents, 

 specimens of the three should be planted in such positions near each other as to ensure 

 that all shall have similar conditions of soil, exposure to sun, wind, and rain, drainage, 

 &c, if the two parents agree in habitat. But in the case of parents which flourish, as a 

 rule, under different environmental surroundings, an imitation of these should be carried 

 out as perfectly as possible under cultivation, or, better still, frequently repeated observa- 

 tions of the wild plants are to be preferred. As proving the necessity for this, the 

 behaviour of Geum intermedium and its parents might be described. G. rivale 

 affects watery places and shady moist copses, always attaining best average growth and 

 ripening the greatest number of fruit clusters when these requirements are fulfilled. 

 G. urbanum grows in " borders of copses, hedgebanks," &c. (Hooker), and thrives best 

 in such circumstances. When grown in a damp place, its vegetative luxuriance checks 

 its reproductive capacity. But this latter situation, if a sheltered one and well exposed 

 to sunlight, hastens its flowering period by a week to a fortnight. For an accurate 

 estimation, therefore, parents growing under normal surroundings must be watched, and 

 the average date of first opening of a blossom on several plants should be taken, while 

 for the hybrid an even wider and more careful record is requisite. 



Geum intermedium. — The first blossoms of G. rivale opened in several places 

 visited this year (1891) on 7th to 9th May ; those of G. intermedium on 20th May in the 

 open, and on the 26th in shady places ; G. urbanum was gathered on 6th June in an 

 exceptionally sheltered spot, but the first flowers opened, in four other localities, from 

 12th to 18th June. 



Carduus Carolorum. — The one parent, C. palustris opened its first capitular florets 

 in several localities from 20th to 22nd June, C. Carolorum on 25th June, and C. hetero- 

 phyllus on 2nd July. 



Erica Watsoni. — E. Tetralix was gathered wild in three places, with first flowers 

 opened, on 24th June ; E. Watsoni flowered in the open beds on the 15th July, and on 

 the Rock Garden at Edinburgh on the 20th July ; E. ciliaris opened in the beds on 

 28th July, and on the Rock Garden on the 14th August. 



During 1887, when the later flowering plants suffered from severe droughts, E. Tetralix 

 blossomed on 30th July, E. Watsoni on 1st August, and E. ciliaris on 14th September 

 at the Rock Garden, but the first named was gathered that year by us in a wild state on 

 the first Saturday of July. 



Rhododendron prcecox. — R. atrovirens opened this year (1891) on 21st January, 

 and from an average of twenty years' observations blossoms on 2nd February ; R. prcecox 



* Trans. Bot. Soc. Edin., vol. xvii., 1888. 



