58 METABOLISM 



There are marked intercellular spaces between these cells, spaces which appear 

 to be always full of water, even though no secretion is taking place. 



In the plants which have been referred to (the Urticaceae and Artocar- 

 paceae, perhaps, excepted) the epithem plays the part of a filter. It is only 

 when a positive pressure has been developed in the vessels that an excretion 

 of water begins, and naturally this excretion takes place at the region 

 of least resistance ; the water leaves the tracheids, enters the interceUular 

 spaces of the epithem, and escapes thence to the exterior. This process is 

 facilitated by the same external factors which aid in the raising of the 

 bleeding-pressure and the exudation of drops, and hence we are able to 

 induce this exudation by elevating the temperature sufficiently, and by 

 moistening the soil and the air, at times when, under normal conditions, the 

 plant exhibits no such phenomena. It has long since been demonstrated 

 that the phenomenon we have been considering is due to the pressure of sap 

 in the vascular system, and that if water be forced into the cut end of 

 a branch by means of a column of mercury, say 20 cm. in height, an 

 excretion of drops may be at once induced. If instead of employing water 

 we use a watery solution of a colouring matter incapable of penetrating the 

 protoplasm (Molisch, 1880), the coloured solution is observed to issue from 

 the leaf teeth unaltered, indicating that the epithem, in so far as it consists 

 of living cells, takes no part in the filtration. Haberlandt (1894) has demon- 

 strated this fact perhaps even more clearly by killing the epithem by means of 

 corrosive sublimate, and showing that the excretion of water continues not- 

 withstanding. On the other hand, he showed that in the Artocarpaceae and 

 Urticaceae the excretion of water ceased on the death of the epithem cells, 

 and for that reason he attributed to them in these cases a special capacity for 

 actively excreting water. This has not, however, been confirmed by other 

 investigators (e. g. Spanjer, 1898), so that at present it may be considered as 

 doubtful whether two types of epithem occur in plants, a filtering epithem and 

 an actively secreting epithem. 



The existence of ' actively ' secreting hydathodes is, however, well estab- 

 lished. Many epidermal cells, i. e. such as grow out into epidermal hairs, are 

 known as hydathodes. Examples of these occur in the hairs found in the hollow 

 leaves of Lathraea (Goebel, 1897, Haberlandt, 1897), and especially in the 

 secretory hairs of many insectivorous plants. The fact that there is, as a rule, 

 no vascular bundle directly connected with these hairs renders it apparently 

 impossible that the excretion of water can take place in these cases by simple 

 filtration. It is obvious that these hairs obtain their water supply osmotically, 

 and that their cells are possessed of an osmotic activity comparable in all respects 

 to that exhibited by those parenchymatous cells of the root which excrete water 

 into the vessels in that organ. The difference is only one of situation, and we will 

 describe as water glands all organs which exhibit unilateral excretion of water. 



Owing to its chemical characters we have previously described the sap, which 

 exudes from the leaf apices of Colocasia, simply as ' water '. Although this no- 

 menclature may be appropriate in the case of Colocasia, it is by no means always 

 so in the case of other hydathodes, whether these be active or passive. Very 

 frequently the fluid contains carbonate of lime, which makes itself apparent 

 in the form of separate crystals or complete incrustations, after the evaporation 

 of the water. Examples are well known to occur in the filtering hydathodes 

 of species of Saxifragaceae, where the occurrence of scales of carbonate of lime 

 in the depressions into which the water filters is well known. The hairs of 

 Lathraea, as already mentioned, are especially good illustrations of actively 

 secreting hydathodes. In these cases the protoplasm must undoubtedly be 

 permeable to lime salts, just as it is permeable in other cases to other salts. 

 Thus in many Tamaricaceae and Plumbaginaceae peculiar glands are found, 

 by the activity of which these plants become covered over with a greyish 



