MORPHOLOGICAL NOTES I97 



the anther of grasses : by this means the shaking out of the 

 pollen is facilitated in all these anemophilous plants. 



Returning now to the morphological question, one might, 

 in arriving at a decision, also take the anatomical facts into 

 account. If (a) (fig. i) represents the bract of the capitulum, 

 then one might expect that the normal orientation would be 

 shown in the constitution of the vascular bundle, phloem on 

 the lower side, xylem on the upper. Of course, I do not 

 regard the anatomical conditions as decisive, since, in my 

 opinion, these are determined by the morphological, not the 

 reverse. In the present instance this would mean that, in the 

 event of a relatively limited development and retardation in 

 time on the part of the bract in comparison with the axillary 

 shoot, it appears quite natural that the former (the bract) 

 should be provided with its vascular system from the latter 

 (the shoot). This being so, one expects the xylem to be 

 uppermost, with the phloem underneath. 



The conditions actually existing are briefly as follows: 

 The thin, much-flattened peduncle of the capitulum contains 

 two vascular bundles (rarely three) ; the phloem-groups of 

 these are set towards the narrow margins, and the xylems 

 are turned towards one another and somewhat obliquely 

 downwards (fig. 3). This anatomical structure is interesting 

 because unusual in a shoot-axis. 



A certain biassed school of anatomists assumes that shoots 

 are distinguished from leaves by their anatomical structure, 

 particularly in the arrangement of the vascular bundles. 

 The shoot of the capitulum of Ambrosia does not diff"er in 

 structure from many leaf-stalks, except that in the latter the 

 xylem of the vascular bundle is directed upwards. 



There is little doubt that this anatomical structure is the 

 result of reduction ; .that is, the peduncle of the primitive 

 type possessed more than two vascular bundles. Linked with 

 these bundles are those which ramify throughout the involucral 

 leaves and the flower, but I have not followed these in detail. 

 It may be mentioned, however, that the vascular bundles of 

 the involucral leaves are normally orientated in that their 

 xylem lies towards the growing-point of the capitulum. The 

 orientation is that described above, and for the reasons given 

 it does not seem to me possible to arrive at any definite con- 



