The compatibility of the birds was assessed based primarily on visual observations of the birds' behavior. This included their diurnal activity patterns, their tendency to disperse from the coop and water source, and most importantly, their feeding habits.

Based upon structured observations, chickens were found to be omnivorous, feeding on insect pests, beneficial insects, weeds, and in some instances, the crops. They were found to feed on two important insect pests, Japanese beetle and Colorado potato beetle, however there was noticeable inter-bird variability in the willingness to feed on the latter pest. Although chickens fed upon vegetation, their effects on weed biomass were small in comparison to the geese. The tendency of the chickens to feed on unearthed potato tubers was generally not a problem during this study because the potatoes were hilled, providing protection from the scratching and pecking which typify chicken foraging. However in later studies in which the potatoes were not hilled, chicken damage to the tubers was a common occurrence. One other problem during this study was the loss of chickens to predators. Loss to predators is probably the most obvious drawback to free-ranging chickens. Because the birds were put into coops at night, the most serious losses to predators were due to hawks, which hunt in the day.

The geese were found to be solely herbivorous. They were observed to feed less often than the chickens but consumed large amounts of vegetation while foraging. The most preferred plants included common dandelion, common plantain, and grasses, but a variety of other species were also consumed. Geese required large quantities of water and spent more time than chickens close to the coop and water source. Geese occasionally fed upon unearthed potato tubers but did not feed on potato or apple foliage. Weed biomass comparisons clearly demonstrated that geese had potential as weed control agents. Furthermore, the geese, which grew faster and larger than the chickens, were less susceptible to hawk predation.

The compatibility study suggested that chickens might have an effect on at least two pests (Japanese beetle and Colorado potato beetle), and possibly others, which spend part of their life cycle on or near the soil surface. By contrast, the geese showed clear promise as weed control agents. Therefore, in these years, the effectiveness of free-range chickens and geese as biological control agents of insect pests and weeds was studied. The ultimate goal was to determine if the presence of free-range chickens or geese provided any crop protection benefits. The effects of the birds on the abundance of four insect pests (plum curculio, apple maggot, Japanese beetle, and Colorado potato beetle), weed growth, potato growth and yield, and apple damage and yield were measured. The direct and indirect ecological effects of the birds, as determined in this study, and the compatibility study, are presented in Figure 1.

Add Figure 1 here

The feeding activities of the geese resulted in clear reductions in weed growth which led to increased potato plant growth. In potato tuber yields in plots weeded by geese were substantially higher than those of control plots. However in the following season, no increase in potato yields were found due to a disease infestation which affected all treatments. Interestingly, weeding by the geese also resulted in greater Colorado potato beetle abundance. Presumably, the presence of large potato plants in a relatively weed-free environment encouraged the colonization and survival of this pest species. Other published research supports this explanation.

A surprising, but consistent, finding was the reduction in plum curculio damage due to the presence of geese. In back to back years, the percentage of fruit with plum curculio damage in plots with geese was less than in control plots. This resulted in a corresponding increase in the percentage of undamaged or pest-free fruit. The compatibility study suggested that the effects of the geese in reducing weeds could result in higher apple fruit yields, however, harvest weights in the plots with geese were not statistically greater than those of the controls in ethir year.

The results of the compatibility study indicated that the chickens could have some effect on weed growth but that densities would have to be increased substantially to achieve adequate control. The chickens showed more potential, however, as insect control agents. Insect pest abundance patterns in both years indicated that the chickens had an effect on Japanese beetle. Furthermore, analyses of chicken gut contents showed that Japanese beetle was a common prey. However, fruit damage assessments showed no differences between the plots with chickens and those without. There are several possible explanations for these contradictory findings: 1) the small plot size allowed for rapid colonization during the period of time between chicken removal from the plots and apple harvest; 2) severely damaged fruit tended to drop from trees resulting in an underestimation of damage in the plots without chickens; and/or 3) wasp damage was mistaken for Japanese beetle damage, causing an overestimation of Japanese beetle damage in some plots.

Plum curculio damage in the plots with chickens was intermediate between the control and goose treatment but did not differ statistically from either of them. If chickens did feed on plum curculio it was probably a relatively rare occurrence and chickens could not be counted on to provide control for this pest. Unfortunately, the effect of chickens on Colorado potato beetle could not be adequately assessed in this study because of the confounding effects of weeds on the abundance of this pest. Further research is needed in this area, particularly to address the effects of intra- and inter-breed variability on the tendency of chickens to feed on this pest.

To assess the compatibility of several common vegetable crops with weeder geese a feeding trial and an unreplicated garden plot study were conducted. The crops tested included tomato, pepper, eggplant, broccoli, cauliflower, cucumber, basil, oregano, and string bean. Plants were 4-6 weeks old at the time of the trial.

The palatability of the crops was initially tested during late June with a feeding trial using five, 10-week-old geese. Each goose was given several common dandelion leaves before being offered leaves of each crop. After the dandelion leaves were consumed several leaves of a given crop were offered. If any of the leaves were consumed by at least one goose the test for that plant was considered positive for palatability.

On 23 June, nine geese were introduced into an experimental garden plot which consisted of two 28 m by 2.5 m beds, each situated between rows of apple trees. The garden beds and tree rows, which occupied 350 m2, were surrounded by electric poultry netting. A total of 223 crop plants were present with each crop being represented by 6 to 61 individual plants. All of the plants were 15-45 cm in height except for the pepper plants which were 7 cm high. Tomatoes and eggplant plants were staked and cucumber plants were trellised; all other plants were without artificial support systems. The geese were maintained in the plot for four days. Each day all of the 223 crop plants were examined for feeding and trampling damage by the geese and classified as having no, light, or heavy damage. A classification of 'light' damage was given if the plant appeared able to recover without yield loss; 'heavy' damage was given if yield loss appeared likely. In addition, weed abundance was monitored by counting all of the weed seedlings in six, arbitrarily-selected 1m2 samples in the garden beds.

The feeding trial indicated that leaves of the two brassica crops were potentially palatable to the geese

See (Table 1).

However, only one of the five geese consumed these leaves and this occurred only after several minutes of masticating it with its bill. During the first two days after the geese were introduced into the garden plot relatively little trampling damage was observed. A thunderstorm which occurred on the third day of the trial resulted in erratic and excited movement by the geese throughout the plot and subsequently caused more damage to the plants. The three crops with support systems, tomato, eggplant, and cucumber, and the shrubby herb, oregano, received no observable trampling damage during the trial. The other five crops, pepper, broccoli, cauliflower, basil, and string bean, had 6 - 22% of plants with light or heavy trampling damage by the end of the trial. Only one crop, string bean, showed any feeding damage by the geese and this was only found on a single plant (Table 1). During the four-day period mean weed seedling densities were reduced from 116/m2 to 40/m2, a 65% reduction.