Transition to the High Mountains of Tropical Africa. 527 
far as the Rocky Mountains of Colorado. Again, in the mountains of 
Mexico we find L. racemosa , growing always at the considerable altitude 
of 3,000-4,500 m. In its narrow leaves, the acute cauline leaves and pendulous 
inflorescences, this plant appears even more similar to the L. spicata than 
L. abyssinica, Parlat. ; it grows also tufted just in the same way as L. spicata. 
But it has mostly rigid leaves, revolute at the edge, and three stamens 
only. 
The facts of distribution and the relations of affinity of the species 
allied to L. spicata lead me to the hypothesis that L. spicata , after having 
originated in the Northern hemisphere, was widely distributed in the 
mountainous parts of it as well as throughout the arctic circle ; that it 
advanced along the Andes of North America as far as Mexico, where it 
was transformed into L. racemosa , and further that from this other peculiar 
species have branched off. It is not only on the mountains connected by 
an arctic-alpine flora during the glacial period that L. spicata was 
distributed, but also on the mountains further south, isolated from the 
continuous arctic-alpine flora. To the east, it appears to have not advanced 
beyond the Himalaya Mountains. When passing over to Abyssinia, only 
few transformations took place : the inflorescence became erect, the basal 
axillary shoots were prolonged like stolons, and the cauline leaves became 
obtuse like the basal leaves : from these characters, Buchenau is inclined to 
believe L. abyssinica , Parlat. ( L . spicata var. simensis ), an hybrid of L. spicata 
and campestris (cf. Engler’s Botan. Jahrb., xii, 130). From Abyssinia to 
Kilimanjaro our plant had to travel a long distance ; but it is not impossible 
that it either still exists or has existed previously on a few of the high 
mountains between Abyssinia and Kenia, from which, having advanced to 
the Kilimanjaro, it again produced new forms not much differentiated up 
to the present. At any rate, it is impossible to do without distribution of 
seeds of alpine plants by air-currents or by birds from one mountain to the 
other in explaining the history of distribution. 
I wish only briefly to mention the two other types of Luziila which 
have reached tropical Africa. L. campestris , (L.) DC., var. Mannii , Buchenau, 
on Fernando Po, at 2,700 m., and on Cameroon Peak, at 3,000-4,300 m., 
shows deviations from the Euro-asiatic forms of L. campestris, somewhat in 
the same direction as L. abyssinica from L. spicata. The growth is taller 
and the leaves are more vigorous. 
L.Johnstonii , Buchenau, also holds about the same relation to L.Forsteri 
as L. abyssinica does to L. spicata ; it has stolons, L. Forsteri not; its 
inflorescence is richer than ordinarily in L. Forsteri ; but specimens from 
Florence and from Tenerife are fully as many-flowered as L. J ohnstonii, 
Buchenau, of the Kilimanjaro, where it occurs in the uppermost region of 
the forest-belt and in the upward extensions of forest between 2,500 
and 2,900 m. 
