,3,3. THE FALLING OF LEAVES. 683 



The result of this process is tlmt the surface of the leaf-scar, 

 directly after the fall of the leaf, is formed, not by the brown dying 

 cell-layers, but by the colourless succulent living cells above these, 

 which, however, rapidly dry up. According to Mohl, the formation 

 of the " separating layer " begins only a short time before the fall ; it 

 was not present on October 4th. It is not formed simultaneously 

 through the whole thickness of the joint, but starts from the inner 

 side bounding the leaf-axil and proceeds towards the outer surface ; 

 it often happens that complete separation has taken place above, next 

 the axil, while the under outer ridge is still firmly united to the 

 cortex and cannot be separated without tearing. He is unable to say 

 at what time the cork-layer was formed, as the researches on Gymnocladus 

 did not extend back beyond the middle of September, when it was 

 already present. The vascular bundles take no part in the changes 

 which anticipate leaf-fall. They run unaltered and uninterrupted 

 from the twig through the cork-layer, and the brown, round-celled 

 layers above it. After the separation in the surrounding tissues they 

 are broken in a purely mechanical way by the mere weight of the 

 blade, or some external strain, such as wind or rain. 



Thus for Gymnocladus Mohl shows conclusively that the separation 

 of the leaf is the result of changes in a distinct "separating layer" 

 which bear no connection with the previously formed cork-layers. 

 The latter serve to protect the surface of the scar, which is conse- 

 quently healed long before it is actually made. 



Mohl also investigated the phenomenon in many other temperate 

 plants with deciduous leaves, including shrubs and herbs as well as 

 trees. In many cases the cork-layer was present as in Gymnocladus, 

 but frequently it was absent ; this was the case in all the ferns with 

 deciduous leaves which he examined, in the beech, several oaks, the 

 elm, the white mulberry, the lilac, belladonna, the tulip tree, and 

 others. Hence, it is evidently not the cause of the leaf-fall as 

 Schacht thought, and as is also seen from the fact that in plants like 

 the poplars, which regularly lose their leaves from below upwards, 

 the cork-layer is in exactly the same condition in those leaves which 

 are about to fall and those still firmly connected ; the difference lying 

 in the " separating layer " which, while scarcely marked in the lower 

 leaves, is completely formed in the upper. The brown coloration 

 (suberisation) of the membranes of the round-celled layer has also no 

 direct influence on the fall, as separation occurs above it in the still 

 fresh and succulent part of the petiole ; this layer varies greatly in 

 size, and is sometimes quite inconspicuous. 



The " separating layer " is universally present, though some- 

 times easily overlooked when only divided from the brown, round- 

 celled layer by two rows of cells, in which case the tear seemed to 

 take place between the dead tissue and the fresh, living leaf-stalk. The 

 formation of this all-important layer usually begins, according to 

 Mohl, a few weeks before the fall, in most cases not before the 



