PECAN" ROSETTE. 15 
muriate of potash, in the other with nitrate of soda. The two limed 
plats free from rosette received in addition cottonseed meal and 
Thomas phosphate, respectively. In the five plats without lime no 
rosette at all developed with the exception of doubtful signs in two 
trees immediately contiguous to a limed plat. The four lime- free 
plats showing no traces of rosette were # the control, untreated, and 
three plats treated respectively with muriate of potash and acid 
phosphate, stable manure alone, and stable manure with ground 
bone. During this period no other cases of rosette developed in the 
vicinity of the experimental block, though cases appeared in other 
parts of this orchard of 700 acres. In two other fertilizer tests where 
the disease was already present at the start, it increased somewhat 
in severity of attack or in the number of new cases in the plats 
receiving lime. 
Analyses of the subsoil around normal pecan trees in parts of an 
orchard free from rosette gave 0.5 to 9.5 per cent of calcium. It 
appears then that the disease is not caused by the presence of lime 
alone, since more lime occurred here than in parts of the orchard 
where rosette was present. Ash analyses of normal and diseased 
leaves and twigs showed only slight or highly variable differences. 
Apparently, however, the percentage of potassium is greater in the 
diseased leaves and twigs. 
In one spray test with Bordeaux mixture on rosetted trees nega- 
tive results were obtained. 
It is evident from numerous orchard records covering periods of 
2 to 12 years that pecan rosette fluctuates from year to year without 
any variation in fertilization or cultural methods. The diseased 
trees may apparently make a complete recovery and remain normal 
for an indefinite period, or after one or more years may again con- 
tract the disease. However, in the majority of cases of recovery 
observed, the trees had not reached the stage where the branches 
were dying back. It seemed thus (56, p. 165, 169) highly probable 
that seasonal climatic changes, such as variations in precipitation or 
moisture content of the soil, might have at least an indirect relation 
to the prevalence of rosette. In large orchards the more or less 
simultaneous appearance of rosette in patches, and its usual limita- 
tion to these areas, suggested some connection with soil phenomena. 
From the apparent nontransmissibility through the seed, negative 
results in attempts at isolation of organisms, the apparently nega- 
tive results of budding and grafting between normal and diseased 
trees, and the results of transplanting tests it was concluded that 
the disease was probably nonparasitic. 
From pruning and " dehorning " tests, transplanting and fertilizer 
experiments, dynamiting of soil around rosetted trees, and from 
orchard records of disease fluctuation it was at that time considered 
