44 Alfred J. Ewart : 



abundantly in the neighbourhood of the Botanic Gardens over 40 

 years ago, and may date much further back still, for its 

 increase is favoured by the conditions attendant on the pre- 

 sence of civilized man. Probably if specimens had been collected 

 from the early part of last century, we vs^ould have received evi- 

 dence of adaptive modification on the part of this plant, but 

 whether R. cruciata, var. australis, is derived from R. bulbo- 

 codium or R. rosea is impossible to say. In any case the whole 

 genus of Romulea is badly in need of revision. 



Experiments on the extermination of this weed are in progress 

 at the Herbarium and in the Domain grounds. The use of pigs 

 has been suggested to root out the corms from the ground, and 

 Mr. T. S. Hall has recorded before the Field Naturalists' Club 

 that white cockatoos coming North from the Otways have 

 performed the sameoffice, and by digging out the corms have cleared 

 patches of ground infested by the weed. Mr. C. French, Jun., 

 reports that he has often seen children eating the bulbs, although 

 to the novice the taste is by no means pleasant. A quantity of 

 the corms crushed and washed through a tine sieve yielded on 

 settling an abundance of fine-quality starch, so that, were it not 

 for the expense of collecting the bulbs, they might form a profit- 

 able source of starch. The seeds are also highly nutritious, and 

 it is owing to the sparrows and other birds which eat the seeds 

 that the plant is so rapidly and widely spread. Future investi- 

 gation may show some use for the seeds. The stalk of the fruit 

 capsule is strongly geophilous, and curves down towards the 

 ground during ripening. In loose ground it is sometimes partially 

 or completely covered before the seeds are shed. 



ScAEVOLA LACINIATA, Bailey. Flora of Queensland. 



This plant resembles the variety pallida of S. microcarpa in the 

 glabrous style and the corolla nearly glabrous outside, but has the 

 larger flowers of the type species and the leaves somewhat more 

 cut. It can only be classed as a variety laciniata, Bailey, of S. 

 microcarpa, intermediate between the type species and variety 

 pallida. 



Scaevola scandens, Bailey, Flora of Queensland, seems to 

 resemble S enantophylla closely in all specific points except 



