10 xMASS. EXPERIMENT STATION BULLETIN 281 



The iiioaii annual precipitation at Amherst, Massachusetts, where these 

 experiments were conducted was 43.49 inches for the period 1889-1928. For the 

 same period the mean total precipitation for May, June, July, August, and 

 September was 19.53 inches, and this amount was rather evenly divided through 

 the growing season. Thus, under average conditions, there is a favorable rainfall 

 during the pasture season in this region. Occasional droughts of three or more 

 weeks' duration have an adverse effect on the production of the pastures located 

 on the lighter, well-drained soil types. 



Experiment A. — The area occupied by this, the first experiment laid out on 

 the Tillson Farm, was on a slight elevation and therefore subject to more leach- 

 ing than most of the area occupied by experiment B. The flora at the time the 

 experiment was started consisted largely of running cinquefoil {Potentilla cana- 

 densis L.) together with smaller amounts of grasses, mosses, and weeds. The 

 area was frequently referred to as the "cinquefoil area." There were five plots 

 in this experiment, four of which received different fertilizer treatments as top- 

 dressing, as shown in table 3. One plot received no fertilizer. The plots were 

 1 rod by 2 rods, and the east half of each plot received limestone in 1921. In 

 1927, on the basis of experience accumulated since 1921, muriate of potash was 

 applied to all plots except the check, and lime to the west half of all plots except 

 the check, in an attempt to improve the fertility of those plots. 



In the first period of this experiment, 1921-26, no quantitative measurements 

 were made of the effects of the fertilizers on yield or character of the flora. How- 

 ever, photographs of certain plots were taken in successive years of 1921, '22, 

 and '23, which speak eloquently of the changes in the character of the vegeta- 

 tion. By referring to figures 3-5 it may be seen that a vegetation consisting largely 

 of cinquefoil was in the course of three years transformed by chemicals to a 

 vegetation composed largely of white clover and grasses. No seed was sown on 

 these plots. It should be noted, however, that 1923 was a good "white-clover 

 year" in this section. 



On account of the larger economic importance which nitrogen fertilizers 

 were assuming, the nitrogen plots were given a double application of this element 

 in the form of Chilean nitrate of soda in 1928. The results of this treatment and 

 the cuiiuilative effect of the previous treatments are shown by figures on yield 

 of two cuttings of grass, as given in table 4. The rainfall was heavy in 1928 

 and grass and clover grew exceptionally well. The area was mowed in 1928 

 but had been pastured in previous years, and it is held that the figures fairly 

 represent the productivity of the respective plots induced under actual pasture 

 conditions. 



While the lack of uniform treatment of the plots through the period 1921-28 

 precludes certain desirable comparisons, some interpretations of value may be 

 made. First, there is the evident response to an application of lime made seven 

 years previous (Plot 3 E vs Plot 3 W). Second, the east half of all plots was 

 definitely superior to the west half, although the west half of all except plot 3 

 received lime in 1927, while the east half received it in 1921 only. The application 

 of 1927 doubtless had some, even if only a slight effect in 1928. It then becomes 

 evident that more than a year may be required for limestone to exert its full 

 effect. The data of table 4 also strikingly show the cumulative effect of potassium 

 and phosphorus. Unfortunately, no deductions can be drawn as to the relative 

 effectiveness of these two elements. Finally, the effect of nitrogen is shown to 

 be marked. In combination with lime, phosphorus, and potassium it has proved 

 very effective. The etTect of these treatments on the protein content of the grass 



