20 Transactions. 



It is a fact of the greatest significance that identical growth-forms are found 

 side by side amongst species belonging to unrelated families. The import- 

 ance of this occurrence is still more emphasized by the fact that other 

 species in far-distant parts of the earth, growing under approximately 

 similar conditions, may likewise possess the same epharmonic forms. That 

 there should be this convergent epharmony, as it is called, seems to lend 

 the strongest support to the view that the effect of an outer stimulus upon 

 the plant, such as light, heat, &c, may become hereditary. 



Only a few characteristic growth-forms receive attention here, and the 

 treatment of these is quite brief. A few others are dealt with when treat- 

 ing of the genus Veronica (p. 44). 



(a.) The Divaricating Shrub Form. 



This very common New Zealand growth-form consists of much -branched 

 often stiff and wiry stems which are pressed closely together or even 

 interlaced, the branching being frequently at almost a right angle (see 

 Plate III, fig. 2). Although I do not know of any example where wind has 

 brought an exact replica of this form, a wind-shorn shrub is closely related. 

 .Still more close is the unstable form assumed by certain lianes in the open 

 (e.g., Rubus, Muehteyibeckia* and Clematis) which grow in company with true 

 divaricating shrubs. Further, the relation to shrubs of an open growth 

 is exhibited by the already mentioned Corokia Cotoneaster and Pittosporum 

 divaricatum, when they grow as forest-plants. Suttonia divaricata Hook. f. 

 (Myrsinac.) is virtually fixed under all circumstances, though in the forest- 

 it may have a slender trunk. 



The ecological factors governing this growth-form appear to be wind, 

 in the first place, and then various other xerophytic stimuli, of which soil 

 must play an important part. 



The most instructive case of convergent epharmony in these plants 

 is in the scrub of certain South Island montane river-terraces or river- 

 beds, where so greatly do many of the species resemble one another that 

 it is quite easy to confuse them. The following is an actual combina- 

 tion : Pittosporum divaricatum Cockayne (Pittosp.), Rubus subpauperatus 

 Cockayne (Rosac), Discaria toumatou Raoul (Rhamnac), Aristotelia fruc- 

 ticosa Hook. f. (Elaeocarp.), Hymenanthera dentata R. Br. var. alpina 

 T. Kirk (Violac), Corokia Cotoneaster Raoul (Cornac), Coprosma propinqua 

 A. Cunn., C. parviflora Hook. f. (Rubiac). Hymenanthera would frequently 

 be absent or confined to specially stony ground. There would also pro- 

 bably be one or more species of Veronica and Carmichaelia, but their 

 growth -forms are different. 



The divaricating growth-form also occurs in the following families : 

 Polygonaceae, Ranunculaceae, Leguminosae, Rutaceae, Icacinaceae, Mal- 

 vaceae, Mysinaceae, and Compositae — i.e., in fifteen families altogether, all 

 of which have likewise members with altogether different growth-forms. 

 Generallv speaking, the earlier juvenile form of these plants is mesophytic. 



(b.) The Cushion Form. 



Every transition exists between the open circular mat-like form and 

 dense unyielding cushions. It is merely a question of degree in reduc- 

 tion of internodes and closeness of growth. The genus Gelmisia shows 



* M . Astoni Pet ric. most closely related to the liane. M. complexa, is a divaricating 

 shrub. 



