EXPERIMENTS WITH PERMANENT PASTURES 23 



terms of grass and clover. However, it should be remembered that nitrogen 

 has been used in this experiment for only two years, the minerals seven years. 

 The question of the effect of applied nitrogen on the presence of white clover 

 in pasture vegetation is one of considerable interest and importance. Although 

 the results of only two years are inadequate for the drawing of final conclusions 

 on this question, they are worth considering as showing trends. Figures of table 

 15 show that in 1929 there were 15 cases in which the percentage of clover was 

 more with the addition of nitrogen, 13 in which it was less, and 2 in which it was 

 the same. At the time of the first clipping in 1930, the cases were 14 in which 

 there was more clover with the nitrogen, 7 in which it was less, and 9 in which it 

 was the same (or no clover). At the time of the second 1930 clipping there were 

 8 cases in which the clover was more, 15 in which it was less, and 7 in which it 

 was the same (or none) with the added nitrogen. Thus this experiment affords 

 only meager evidence of a trend toward reduction in stand of white clover due 

 to the application of nitrogen, and no conclusion can be drawn. The evidence at 

 hand indicates, however, that white clover will maintain itself better in the 

 presence of applied nitrogen if there is also a generous supply of calcium, phos- 

 phorus, and potassium. 



The dual effect of nitrogen and potassium on the nitrogen content of the 

 herbage is well brought out by the data of table 19. In plots where the percent- 

 age of clover was low and grass high, the added fertilizer nitrogen almost inva- 

 riably caused an increase of nitrogen (figured as protein in table 19) in the crop. 

 On the other hand, potassium in connection with lime caused an increase in the 

 percentage of white clover, and where this plant was abundant, the percentage 

 of protein ran high — sometimes much higher on the portion of the plot with min- 

 erals only than on that portion which received minerals and nitrogen. Thus the 

 percentage of protein in the crop was a resultant of these two, and probably 

 other factors. The interaction of the two factors caused some interestitig results. 

 For example, in plot 1 there was practically no clover, and the percentage of 

 nitrogen in the grass and weeds was almost doubled by the application of nitro- 

 gen; while on plot 7 the relationship in respect to nitrogen content was prac- 

 tically reversed because of the presence of more clover in the portion of the 

 plot which received only minerals. On plot 22 the clover was abundant in both 

 portions of the plot and consequently there was a high protein content in the 

 vegetation from both sides, it being a little higher where nitrogen was applied. 

 Table 19 further shows, however, that although the percentage of protein was 

 sometimes higher in the crop which did not receive nitrogen than in that which 

 did, the total pounds of protein per acre were invariably higher where nitrogen 

 was used because more pounds of dry matter were produced. These examples 

 indicate the importance of data on botanical composition in the interpretation 

 of pasture experiments of this type. 



That the full effect of nitrogen is not always sliown by the figures on yield 

 of dry matter is further shown by the ratio comparisons in tables 18 and 19. 

 With few exceptions, all of which were due to the greater percentage of clover in 

 the no-nitrogen portion, added nitrogen caused a higher ratio of protein than of 

 dry matter in the plots receiving minerals only and nnnerals i)lus nitrogen 

 respectively. 



From experiment C was first obtained evidence (3) of the toxicity of sodium 

 nitrate for a species of moss {Polytrichum commune L). This moss is common on 

 Massachusetts upland pastures which have become depleted in plant nutrients. 



