and of some allied Scrophulariaceae . 393 
point out that the feeble excretion of water goes side by 
side with the numerical weakness in dome-shaped glands. 
Conclusions. 
So far I have shown that water is excreted in a liquid form 
by the leaves of Lathraea , Pedicidaris , Rhinanthus , and 
Odontites , all of which possess dome-shaped glands on their 
lower faces. The following facts tend to prove that the 
water is poured out by the dome-shaped glands : — 
1. The amount of water poured out is large when these 
glands are numerous (Lathraea, Pedicidaris ), and small when 
they are few in number (. Rhinanthus , Odontites). 
2. There are indications that only those portions of the 
leaf-surface which bear these glands excrete water (Lathraea, 
Rhinanthus). 
3. The cuticle of these glands, in the three cases investi- 
gated, is perforated by a pore \Lathraea (1), Pedicular is, 
Rhinanthus ] which renders that region of the wall permeable 
to water. I therefore regard this pore as a water-way, not 
as a channel through which mucilage may be excreted, as 
is generally supposed. 
4. There is a close connexion between the tracheides in 
the fine bundles and these glands. I can confirm the accounts 
given by Scherffel (1), who alone of previous observers has 
correctly described the structure of these glands in Lathraea 
and their relations to the tracheides. As Scherffel points out, 
it is not true that a bundle of tracheides terminates under 
each gland. I find a network of tracheides very close beneath 
the epidermis lining the pocket. At certain points, especially 
at nodes of the network, certain broad ampullate tracheides 
bulge out towards the epidermis ; but these lateral bulgings 
are less numerous than the dome-like glands, and, moreover, 
capitate hairs stand over them as well as dome-like glands. 
There is no trace of the cuticularization of the walls of sub- 
jacent parenchyma-cells where they bound intercellular spaces. 
In the plane scales on the floral axis of Lathraea , the dome- 
