Problem of Xeromorphy in Marsh Plants . 821 
additional bands r mining inwards from the margin , betzveen the main veins 
(PI. LXXXII, leaves 4, 6, 9; also PI. LXXXIII, leaf 3). 1 The partly 
hairy leaves on the erect flowering shoots also follow the same rules 
(PI. LXXXI, leaves 6, 7, and 8). All the leaves photographed (PI. LXXXI- 
LXXXIII) are mature. The hairs would not spread subsequently from 
the margin over the entire surface. 
Broadly speaking, the rule followed in every case is that when a leaf or 
leaflet is only partially hairy, the hairs are localized principally at those 
points which are most remote from the main water supply. This point will 
be dealt with later. 
§ 6 . Comparison with other Hairy Plants. 
Spiraea Ulmaria is not unique in this curious periodic production of 
hairs, though it is the most striking case I have seen. Specimens 
of X. vestita , Wall., 2 a species quite distinct from X. Ulmaria , though closely 
allied to it, were examined in the Cambridge University and Kew Herbaria. 
The upper leaves are downy, but in both herbaria specimens exist which 
show that the lower leaves are only partly hairy. In one or two cases the 
lowest leaf of all is glabrous. Thus this species apparently follows much 
the same rules as X. Ulmaria. I have also grown Spiraea tomentosa , L., 
from specimens kindly sent me by Professor H. C. Cowles, of Chicago, and 
Professor J. B. Pollock, of Michigan. This is a shrubby form, with small 
downy leaves. Here, too, as the buds open in the spring-time, there 
is a similar succession of glabrous, partly hairy, and hairy leaves. 
A number of other marsh herbs resemble these species of Spiraea in 
forming small glabrous leaves in spring-time, followed, as the season ad- 
vances and the erect flowering shoots are formed, by a steadily increasing 
production of hairs on the later leaves. 3 Such are Epilobium hirsutum , 
Lysimachia vulgaris , 4 Mentha aquatic a, Valeriana sambucifolia , Scabiosa 
succisa, Eupatorium cannabinum , Lycopus europaeus , &c. 
But the phenomenon is not confined to marsh plants. I noticed 
a specimen of Rttbus discolor , W. and N., growing partly shaded by trees in 
the New Forest, in which the proximal leaves on the lateral shoots were 
invariably much less hairy than the distal ones. Again, I examined the 
cultivated Raspberry ( Rubus idaeus , L.), without finding any trace of 
1 The distribution is not always quite so regular as shown in the series of leaves photographed ; 
though it is so in most cases. Sometimes an inter veinal band is incomplete : in this case its inner- 
most portion is often present as a group of hairs in a fork where two veins meet (PI. LXXXIII, leaf 2). 
It will be shown later that hairs are developed also on the main veins themselves. Cf. Renner (’ 08 ), 
p. 128, who states that in many leaves in which the lamina is folded in the bud, both on the midrib 
and lateral veins, hairs are developed especially on the veins and the edges of the leaf, e. g. Fagus 
sylvatica and Carpinus betulus. 
2 S. vestita , Wall. MSS. = S . Kamtschatica , var. himalensis , Lindl. Cf. Lindley (’ 41 ), t. iv, 
and Hooker (’- 79 ), p. 323. 
3 Yapp (’ 08 ), p. 691. 
4 Yapp (’ 09 ), Text-fig. 1, p. 278, 
