BY H. A. LONGMAN AND C. T. WHITE. 163^ 



free and stand erect, usually one on each side of the central 

 style, which may still be looped in. In still later stages 

 they tend to shrivel and are easity misplaced. In no case 

 was any enlargement to be detected at the base. The 

 differences between the normal and abnormal flowers are 

 Illustrated in text-figure 22. 



The flowers of our Proteacese are tj'pically protandrous, 

 and the remarkable mechanism which ensures cross fertil- 

 isation — the immature stj'le when detached from the 

 anthers being a temporar}^ depot for pollen grains — has been 

 described by Delpino, Bentham and a number of subsequent 

 authors. As long ago as 1882, W. Trelease made interest- 

 ing comments on the various views held*. 



In Buckinghamia celsissima the central style may have 

 a number of f)olIen grains adhering to it when first set free. 

 Pollen grains also adhered readily to the supple mentar}^ 

 style-like processes when artificially supplied, but no grains 

 were actually detected upon them in a state of nature. 

 It is, however, conceivable that these processes may be of 

 value as pollen carriers through grains falling upon them, 

 and a still further development would increase their 

 utility. On the other hand, this variation ma}' be of use 

 from the decorative standpoint to the tree, which attracts 

 large numbers of insects, principally Hymenoptera. 



Upon inquiry we foiuid that the tree examined had been 

 brought as a seedling from another specimen at Enoggera. 

 Manj^ racemes from the parent tree were then observed 

 and also a single raceme from a young one close by. With- 

 out exception, each flower showed a pair of supplementary 

 style-like processes developed similarly to those on the 

 first tree. 



Searching also revealed one or two flowers with, 

 elongate processes in each of many racemes examined from 

 trees in the Brisbane Botanic Gardens. In each raceme 

 from another tree at Indooroopilly a few flowers were found 

 with two elongate processes and a number with only one 

 elongate process, the great majority being normal. 



*Prob. Boston Socy. Nat. Hist., xxi, 1882, p. 419. 



