Guthrie- Smith. — Grasses of Tutira. 515 



open land, and that the whole countryside was forest, fern, 

 flax, and raupo. Probably, however, several of the varieties 

 that have now spread, or become noticeable, may have pre- 

 cariously survived on cliffs, shingle - beds, sandspits, and rocks 

 — any spot, in fact, where they had not been smothered by the 

 enormous growth of fern. 



The most efficient plan in observing these natives is to 

 mark a particular plant in, say, a cutting well above reach of 

 sheep, or in some barren gully in a lightly stocked portion of 

 the run — anywhere, in fact, where the grass is likely to be 

 undisturbed for several seasons. There, in its self-chosen spot, 

 the natural habits of the plant can be correctly determined, 

 and this is the method I have followed for many years. 



Last year, however, I thought a native-grass garden would 

 give even more accurate results. What with weeds, however, 

 and abnormal growth, the outcome was a total failure, and I 

 found that in fertile land and a soil constantly stirred to keep 

 down weeds the plant's true characteristics were lost. With 

 the utmost care, and dealing even with minute fragments of 

 sod, it was impossible to eliminate the seeds of other grasses — 

 even their roots in some cases — and there was a constant insur- 

 rection of these unwanted grasses, besides the growth from seed 

 of sorrel, sow-thistle, thistle, and other weeds, all of which 

 must have been lying dormant among the roots of the particular 

 species selected for observation. Then, the growth was very 

 misleading for practical purposes, Agropyrum scabrum reaching 

 7 ft. 6 in. when held up straight, and Dichelachne crinita standing 

 erect 5 ft. 6 in., and with seed-stems like fine-grown straw. 



The twenty-one native grasses on Tutira are : Isachne 

 australis, Microlcena stipoides, Microlcena avenacea, Hierochloe 

 redolens, Echinopogon ovatus, Deyeuxia Forsteri, Deyeuxia quad- 

 riseta, Dichelachne crinita and var. intermedia, Deschampsia 

 ccsspitosa, Trisetum antarcticum and slender form, Danthonia 

 semiannularis, Danthonia pilosa, Arundo conspicua, Arundo 

 fulvida, Poa anceps, Poa ccespitosa, Poa Colensoi, Poa imbecilla, 

 Agropyrum multiflorum, Agropyrum scabrum, Asperella gracilis. 



Isachne australis grows only on the marshy edges of the lake, 

 is of no particular value, and does not seem to spread. 



Microlcena stipoides, in 1882, was the most widely spread of 

 my native grasses, and since then it has fully held its ground. 

 Its value is dependent on its treatment, and when newly burnt 

 it is a nutritive grass, and is kept closely cropped. As it becomes 

 rank, however, it becomes less and less palatable, until at length 

 sheep will sooner starve than touch it. When growing amongst 

 manuka the stems add season by season to their growth, until 

 the grass has almost developed a climbing habit. I have 



