164 MUTATION IN A PROTEACEOUS TREE. 



The first three trees demonstrate a race with well- 

 •developed supplementary style-like processes. There is 

 no evidence of a graduated change from the tiny segments 

 of the h3^pogynous gland, and it is, therefore, thought 

 that this marked modification is better expressed as a 

 mutation than as a variation. With the occasional except- 

 ion in HaJci^a digyna Ewart and Davies* (through fusion 

 of pedicels), a one-celled ovary seems to be the prevailing 

 condition in Proteaceae, and we do not know sufficient of 

 their ancestry to warrant an interpretation of these structures 

 as instancing a primitive tri-locular form. It has been 

 ■suggested that the gynaecium of Proteacese is bi-carpellary 

 because bidented or bifurcated style-ends are found in the 

 genera Agastachys and Adenanthos, but Bentham expressed 

 Tiis opinion that these genera " are essentially monocar- 

 pellary.'^t 



The hypog^aious gland or glands usually possess 

 fairly constant generic characters, and their value in classi- 

 fication was noted by Robert Brown. J He also recorded 

 their resemblance to stamina and their function as secreting 

 organs (loc. cit., p. 133). 



Instances of twin styles in Proteacese have been noted 

 by Cheel {Grevillea punicea R. Br.) ft and Fletcher 

 (Grevillea buxifoUa R. Br.),J{ as well as in Hakea digyna 

 Ewart and Davies.** 



A normal raceme of Buckinghamia celsissima is well 

 illustrated in Bailey's Queensland Flora (Vol. iv, p. 1352, 

 pi. Ix). 



♦Flora of the Northern Territory, 1917, p. 85 

 tJoum. Linn. Socy., Bot., xiii, 1873, p. 63. 

 JMiscel. Bot. Works, Ray. Socy., ii, 1867, p. 17. 

 ttProc Linn. Soc. N.S.W., xxxvi, 1911, p. 158. 

 tJProc. Linn. Soc. N.S.W., xlii, 1917, p. 247. 

 **Flora of the Northern Territory, 1917, p. 85, 



