274 ROUND THE YEAR 



increase, but cannot possibly diminish, the strength of 

 the rod. Actual trial, however, proves that this very 

 natural supposition is wrong. The cased rod, ringed 

 in one place, will not stand vigorous and sudden 

 bending, but will snap across at the ring. Rigidity 

 everywhere but in one place is highly unfavourable 

 to that nearly uniform curvature which enables the 

 rod to endure a bending strain without fracture. All 

 the bending is now concentrated upon one place 

 instead of being distributed throughout the whole 

 length of the rod. Mechanical engineers have long 

 recognised that abrupt changes of section greatly 

 increase the tendency of axles and shafts to break 

 across. 



The separating layer will therefore weaken the leaf- 

 stalk, and predispose it to part at one particular place, 

 although it does not pass through the vessels and 

 fibres, which are the principal means of attachment of 

 the leaf. But this is not all. The separating layer 

 contains a mechanism for producing a positive thrust, 

 which comes in aid of the pull of gravity and wind- 

 pressure, and suffices to part the leaf from the branch, 

 even when it is supported and kept in a perfectly still 

 atmosphere. How this i-s accomplished I shall try to 

 explain a little later on. 



The fall of the leaf is an old subject of enquiry, but 

 the material facts necessary to a satisfactory ex- 

 planation were not discovered till the year 1859. It 

 happened in that year that the eminent botanist, Hugo 

 von Mohl, spent his long vacation at home instead of 

 at the seaside or in the mountains, and was thus led 

 to observe the fall of the leaf with all the advantages 



