46 EOLAND M. HARPER 



tion must be dissolved in rain-water, especially where the vege- 

 tation is subject to fire, and carried off into streams.) In my 

 samples the ash percentages ranged from less than 3 in Typha 

 latifolia to over 11 in the upland meadow, which occupies the 

 most fertile soil and makes the best hay. The ash of my 

 Phragmites and Typha angustifolia weighed more than the en- 

 tire hay in some of the western localities. 



The ash samples themselves present some interesting differ- 

 ences. Those of Spartina and Typha latifolia after the final 

 burning consist largely of hard porous lumps strongly resembling 

 the slag from an iron furnace, and perhaps containing consider- 

 able iron and phosphorus, like the slag, as well as sodium from 

 the salt water. That of Typha angustifolia is gritty, some- 

 thing like old plaster, while the remainder are fine gray powder 

 like ordinary wood ashes, probably consisting largely of potas- 

 sium carbonate. An analysis of these ash samples has not yet 

 been possible, but it would doubtless give some interesting 

 results. 



These few examples, while perhaps of little value in themselves, 

 may serve to point the way to an intieresting and almost un- 

 trodden field of investigation. Similar studies of herbaceous 

 vegetation should be made in other parts of the world by those 

 who have the opportunity and the necessary enthusiasm. The 

 method might be extended to shrubby vegetation, like the heaths 

 of the North, the scrub of Florida, and the chaparral of Cali- 

 fornia, by cutting clean an area of a few square yards without 

 disturbing the roots, and then weighing the new growth on a part 

 of it a year or two later. It will probably be found that similar 

 combinations of soil and climate will in the long run produce 

 about the same amount of vegetation per unit area per annum, 

 whether it is herbs, shrubs or trees. ^ 



* It may not be generally known to botanists that the reports of the last few 

 United States censuses give the acreage and tonnage of "wild, salt or prairie 

 grasses," cut for hay in each state and county, from which one can get an idea of 

 the extent of prairie, meadow, or marsh, and the rate of growth of the vegetation 

 in different regions. The average yield of such hay is about a ton to the acre, and 

 the maximum about 3 tons; but the figures should not be taken too literally. 



