Transition to the High Mountains of Tropical Africa. 539 
a few more cases of disjointed distribution which seem to be the result 
of some other evolution than that of the species considered so far. 
One of the most prominent instances of disjointed range is presented 
by the genus Canarina. For a long time, only one species was known, 
C. campanula , Lam., a plant very characteristic by its large, bell-shaped, 
orange-coloured flowers, On the Canary Islands, it is very frequent in 
some places ; it is often cultivated in greenhouses. It was considered to be 
an isolated type of the Campanulaceae, an endemic product of the Canary 
Islands. But quite recently, a second species was discovered by Dr. Stuhl- 
mann on the Emin Pasha expedition : C. Etninii , Ascherson (in Sitzber. 
naturf. Freunde, Berlin, 189a), which was found in the forest-region of 
M. Ruwenzori, at 3,500 m. a. s. 1 . A third species, C. abyssinica, more 
allied to the second one, was collected by Dr. Ellenbeck, who accompanied 
as the physician the expedition of Baron Carl von Erlanger to Somaliland 
and Gallaland. It was found in Galla Highland, west of Lake Abbaya, at 
2,000-2,500 m. a. s. 1 ., later on also more towards the west by O. Neumann. 
These plants, having berries eaten by men on the Canary Islands and likely 
to be consumed by birds also, may be assumed to have been spread from 
mountain to mountain by birds. The rather striking differences, however, 
between the Canarian species and the two species of the African continent 
seem to indicate that Canarina is an older genus whose species, having 
travelled in more remote periods, may have had wider areas formerly. 
It is not impossible that this genus was indigenous even to the Mediterranean 
region at some distant period. I am inclined to the consideration of this 
hypothesis by the distribution of Sempervivum arboreum , which is spread 
throughout the southern Mediterranean countries, from Portugal and Spain 
as far as Cyprus. By natural affinity it is connected with S. chrysanthum , 
Hochst., a native of Abyssinia, and at the same time with a number 
of species (viz. N. canariense , L., N. urbicum , Chr. Smith, S.palmense , (Webb) 
Christ, S', cuneatum , (Webb) Christ, N. ciliatum , Willd., N. percarneum , 
Murray, S', tabulaeforme , Haw.) which are indigenous to the Canary Islands 
or (S. glandulosum , Ait., S. glutinosum , Ait.) which are natives of Madeira. 
With regard to the species or varieties I have enumerated above 
(Canarina excluded), I may be allowed to make some general remarks. 
We may call such modifications CLIMATIC AL ADAPTATIONS, but only in 
this sense, that this adaptation is a passive one, caused by the physical 
conditions of the climate, not an active one, which would correspond to 
the views of Lamarckians. 
The first condition for such modifications is that the fruits or seeds 
are such that, without losing the faculty of germination, they can be easily 
transported to regions removed from the locality where they were indi- 
genous before. 
This will in many cases happen accidentally. It is also only accidental, 
pp 
