136 PLANTS WITH TRAPS AND PITFALLS TO ENSNARE ANIMALS. 



are broadly cordate, and they give the impression of being mounted fairly and 

 squarely upon the stem by means of the highly swollen and notched basal portion. 

 But it is only necessary to detach one of the scales from the stem to convince 

 one's self that this is not the case, and that the part taken at first sight to be the 

 underside or back of the leaf is only a portion of the superior surface. In reality 

 each of these thick squamiforui leaves is rolled back, and in it the following parts 

 may be distinguished: first, the place of insertion on the stem (fig. 25 3 ), which is 

 relatively small; secondly, the portion taken on cursory examination to be the 

 whole upper surface of the leaf, and consisting of an obliquely ascending blade 

 limited by a sharp border; next, starting from this sharp border, the part which* 

 owing to its being suddenly bent down at an acute angle and falling away steeply, 

 is usually taken for the dorsal or inferior surface of the leaf, but which belongs, in 

 point of fact, to the front of the lamina; fourthly, the free extremity of the leaf in 

 the form of an involute limb; and fifthly, the true dorsal part, which is very small 

 relatively and is not visible until the involute tip is removed. Owing to the 

 involution of the apex, a canal or rather a recess is formed and runs across beneath 

 the leaf, close under the place where the latter is joined to the stem (see fig. 25 2 ). 

 From five to thirteen (usually ten) chambers open into these recesses through a 

 series of little holes. They are excavations in the thickness of the scales and are 

 probably, in this form at any rate, unique in the realm of plants. These extraordi- 

 nary chambers must be described as deep excavations in the foliar substance 

 proceeding from the back of the leaf. With a view to elucidating their function 

 in relation to the life of the plant, their structure must be more particularly 

 considered. 



The chambers radiate as it were from the orifice at the base of the leaf. Though 

 closely adjoining one another, they are not in lateral connection by means of pass- 

 ages or canals. Their walls are irregular and undulating (see fig. 25 3 ), and are 

 characterized by the peculiar structures which are borne on the lining raised up 

 above the ordinary epidermal cells and projecting into the cavity. These structures, 

 of two sorts, are shown in fig. 25 4 , under a considerable magnification. One sort, 

 and these are by far the more numerous, are of the nature of short capitate hairs. 

 The head is formed of a pair of cells, and they are supported on a short cylindrical 

 cell which serves as a stalk. The other sort is sparsely scattered amongst these 

 capitate hairs. They are oval in outline and but slightly raised above the ordinary 

 epidermal cells. Each consists of a tabular cell upon which rests a slightly convex 

 cushion composed of not more than four cells all lying in the same plane. One such 

 sessile gland is shown in the centre of fig. 25 4 . In this case the cushion consists of 

 three cells. A further peculiarity has been observed in these sessile glands. The 

 summit of each is marked by a tiny pore (not shown in the figure), an actual hole 

 in the wall at the geometrical centre of the convex surface. 



In the wall of the chamber, just below the lining epidermis, run the vascuL 

 strands (fig. 25 3 ). The vessels of which they are composed form a considerable 

 plexus or net-work in this region. Now it is known that the ground in which 



