Ash Flower Gall: Within Tree Distribution and Chemical Management

Ash Flower Gall Sampling

Gall density estimates for each tree obtained on 1 and 8 July 1988. The crown of each tree was visually divided into thirds (top, middle, bottom). Three branches were randomly chosen from each crown level for a total of nine terminals per tree. The terminal 60 cm of each branch comprised the sample unit from which all current season’s galls were counted and recorded. Previous season’s (1987) galls were brown-black in color and easily distinguished from the current year’s (1988) green galls.

Counts of galls per crown level were transformed using natural log (x + 1) and analyzed by 3-way analysis of variation (ANOVA). Means were separated using Tukey’s honest significant difference (HSD), (7).

Insecticide Efficacy Testing

Treatments were applied on 18 April (prior to ash staminate flower bud break) and on 10 May 1988 (ash staminate flower full bloom stage) when winds were light and variable. Treatments were evaluated as above on 1 and 8 July 1988. Experimental treatments were: dicofol (Kelthane) 35WP, fluvalinate (Mavrik Aquaflow) 22.3% flowable, insecticidal soap (Safer), carbaryl (Sevin) 80S, carbaryl (Sevin) 80S + insecticidal soap (Safer) and an untreated control (UTC). Trees were sprayed to run-off by a professional applicator.

The 10 May 1988 application differed slightly in procedure. Carbaryl was not sprayed on this date, and dicofol was applied to a different (previously untreated) tree in each replicate.

Post-treatment counts of galls per tree were transformed using natural log (x + 1) and analyzed by 2-way analysis of variation (ANOVA). Means were separated using Tukey’s honest significant difference (HSD).

Sampling/Crown Distribution

Significant differences existed in mean gall density among the three crown levels over the 63 trees sampled (Table 1). Approximately 62%, 25% and 13% of the galls were contained in the top, middle and bottom crown levels, respectively. However, there was an interaction between height and replication (P = .047) suggesting that gall distribution can vary from tree to tree. In seven of nine replicates, galls in the top crown level outnumbered the other two levels, while middle and bottom crown levels were more variable.

An accurate assessment of gall density will require top crown level sampling. However, municipalities, or other operations involved in large scale spray programs for AFG, could reduce sampling time by first examining the lower crown level. Those trees having unacceptably high gall numbers in the lower crown will not require additional sampling since the middle and top crown levels will have at least as many galls. Only those trees with tolerable gall numbers in the lower crown would require upper crown sampling.

It is possible that the chemicals applied to the trees in this study may have influenced gall distributions. However, the results show that trees in the UTC, which should represent naturally occurring distributions, did not differ significantly from the treated trees with respect to mean gall distribution (Table 2). Apparently, in this study, chemical treatments did not significantly affect gall density and distribution within trees. Therefore, data reported here should accurately reflect natural conditions.

Insecticide Trials

Significant differences in efficacy existed among the seven treatments, however, none of the treatments differed significantly from the untreated control (Table 2). Trees treated with dicofol 35WP applied on 10 May 1988 and fluvalinate applied on 18 April and 10 May 1988 had more galls than the UTC trees. The carbaryl 80S treatment applied on 18 April 1988, had the lowest gall numbers (Table 2).

Carbaryl 80S appears to be the most promising treatment. This is interesting since it was only applied once compared with fluvalinate, insecticidal soap and the carbaryl/insecticidal soap mix which were all applied twice. In addition, trees treated with dicofol 35WP on 18 April 1988 had significantly lower gall numbers than trees treated with dicofol 35WP on 10 May 1988, also indicating that a single effective treatment prior to bud break may provide sufficient control.

Dicofol 35WP (the previous standard for AFG control in the tree care industry) applied at flower full bloom stage and fluvalinate (applied on both dates) resulted in higher gall numbers. Predator mites of unknown species were consistently observed on flowers at full bloom stage. Therefore, it is possible that predators of E. fraxiniflora may be most active at flower full bloom stage and susceptible to these chemicals. However, carbaryl 80S would also have to be tested at full bloom stage to accurately assess this assumption. In addition, the carbaryl/insecticidal soap treatment (applied on both dates) resulted in lower gall numbers as compared with the other treatments. The lower gall numbers may indicate that carbaryl is not affecting the E. fraxiniflora predators which were observed or that the first application reduced E. fraxiniflora numbers enough that predator mortality was insignificant.

The 1988 season was the second consecutive year that E. fraxiniflora was inadequately controlled (Wawrzynski, unpublished data) using standard treatments applied at the recommended times (2, 9). This suggests that the life history of E. fraxiniflora as briefly discussed by Felt (3), Connold (1), Felt and Rankin (4), Garcia (5), and Wawrzynski and Ascerno (9) may need further study.

Overwintering mites were observed approximately three weeks prior to bud break. Because the mites are active this early, it is possible that they have penetrated beneath the flower bud scales before the recommended treatment time of flower bud break (2, 9). To be most effective, a material with sufficient chemical residue should be applied a few weeks prior to bud break. Dormant oil application, as suggested by Felt and Rankin (4) may be the most effective treatment.

Ash flower gall, although a very common problem, may not be as well understood as thought. Lack of previously published information on AFG distribution in tree crowns, has limited options for more efficient control. It is obvious from these data that more attention needs to be given to spray applications in top crown portions of the tree. In addition, if AFG is an aesthetic problem, gall distribution could be used to begin aesthetic injury level studies. A better understanding of public perceptions of AFG and how these relate to gall distribution, could be used to influence control measures. Finally, chemical controls were shown to be erratic and mostly ineffective. Increased study of E. fraxiniflora is more likely to yield information useful for control than screening new chemicals.