1 68 Journal of Agricultural Research [vo1.iii,no.2 



RELATION TO THE SOIL 



The relation of pecan rosette to the soil is partially elucidated by the 

 following experimental data. 



Severe pruning of rosetted trees or cutting back to a stump gave nega- 

 tive results (see p. 152). The new growth was often healthy in appear- 

 ance, but toward autumn or during the following season symptoms of 

 rosette invariably reappeared (PI. XXVIII, fig. 2). The fact that in no 

 case was the disease eliminated by this treatment is at least not unfavor- 

 able to the view that it is contracted directly or indirectly through the soil. 



Transplanting of healthy nursery trees to holes from which rosetted 

 trees had been removed gave a very high percentage of rosette in the 

 replants (see p. 152). These results point toward the conclusion that 

 some relation exists between rosette and the soil, either directly through 

 the soil itself or through previous infection of the soil by rosetted trees. 



The transplanting of badly rosetted trees to situations where less 

 rosette or none at all has been observed gave a very high percentage of 

 recovery (see p. 153). All rosetted trees replanted at Washington, D. C, 

 recovered, though many of them were dying back with the disease when 

 first taken up (see p. 154). In the latter cases the entire change of both 

 soil and climate is held accountable for the uniform recovery of all 

 rosetted trees. All these transplanting tests with rosetted trees tend to 

 indicate the soil relation as the direct cause rather than infection of the 

 soil with a parasitic organism or virus from previously diseased trees. 



The results of the fertilizer experiments were partially negative (see 

 p. 159). However, in a i6-plot test upon normal trees at Baconton, Ga., 

 9 out of II limed plots developed cases of rosette, while i out of 5 unlimed 

 plots showed doubtful traces on two trees bordering a limed plot. Care- 

 ful observation at the beginning of the experiment did not reveal a single 

 case of rosette, and at the last observation no rosette was found in con- 

 tiguous parts of the orchard. The plots receiving lime, acid phosphate, 

 and nitrate of soda and those receiving lime, acid phosphate, and muriate 

 of potash developed by far the most rosette. In fertilizer experiments at 

 two other points considerably more rosette developed on limed than on 

 unlimed plots. These results, while not in all cases very definite, have at 

 least tended to show that some relation exists between rosette and the 

 constituents of the soil and that the lime content alone or in combination 

 with other substances may have a varying causative effect. 



Dynamiting the subsoil in two different orchards gave negative results 

 (see pp. 161-162). This fact, together with observations showing the 

 absence of rosette in swampy land (see p. 150), seems to indicate that the 

 disease is not directly due to lack of proper subsoil drainage. 



Orchard records of individual trees over periods of 2 to 12 years show 

 considerable fluctuations in the disease, irrespective of treatment and 

 without any special treatment (see p. 163). Trees with a mild or mod- 

 erate attack frequently recovered, and even when the staghorn stage 



