m WW Wu £th 12 ‘ M 400—1 I—so \ (I) THE EFFECT OF LAWN AREA 0N BREEDING DENSITlES AND REPRODUCTEVE OUTPUT IN AMERRAN ROB!NS 00-h Thesis for the Degree of M. S. MECHEGAR STAYS UNIVERSITY MECHAEL SEAN CHAMPAGNE 1.975 JW/lfl/H/Qfl/WMWWW 51/ '“ ““ ‘" ; LIBRARY Michigan §ta4t>a :: -« Umcrszry ABSTRACT THE EFFECT OF LANN AREA ON BREEDING DENSITIES AND REPRODUCTIVE OUTPUT IN AMERICAN ROBINS By Michael Dean Champagne In the spring and summer of T974 and 1975 American robins were studied in residential and nonresidential locations in Lansing and East Lansing, Ingham County, Michigan, in order to investigate whether percentage lawn area is associated with nesting density and success. Nine study plots were chosen for study during 1974 and seven study plots were chosen for study during l975. Plots exhibited a range of lawn areas from 44 percent to lOO percent in l974 and 46 percent to 100 percent in l975. The number of attempted nests (nests that reached the egg stage) was used as an indicator of relative robin densities among study plots of different percentage lawn area. The number of successful nests (nests that fledged young), total young fledged, and young fledged per successful nest were used as indices of reproductive success in the study plots. In general, density and reproductive success decreased with a decrease in percentage lawn area in l974. In l975, density and reproductive success did not decrease with a decrease in percentage lawn area. When combined, the l974 and 1975 results indicate that density and reproductive success did not decrease with a decrease in percentage lawn area. The 1974 results seem to indicate that some correlate of percentage lawn area, possibly food, was acting as a limiting factor on robin Michael Dean Champagne densities and reproductive output. The 1975 results indicate, con- versely, that density and reproductive success were not limited by a correlate of percentage lawn area. Two possible explanations for the results are presented. First, it may be the case that limiting factors do not operate in every year, thus, the contrasting results of 1974 and 1975. Second, correlates of percentage lawn area may not be limiting at all. That is, the statis- tically more satisfying results of 1975 may be indicative of the real relationship between robin p0pu1ations and percent lawn area, and the results of 1974 may represent a random occurrence, intensified by the tendency of robins to be locally oriented. THE EFFECT OF LAWN AREA ON BREEDING DENSITIES AND REPRODUCTIVE OUTPUT IN AMERICAN ROBINS By Michael Dean Champagne A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Zoology 1975 ACKNOWLEDGMENTS I wish to thank the many pe0ple who c00perated with me on this project including the Lansing and East Lansing city planners who helped me obtain aerial photographs, the groundskeepers of Forest Akers Golf Course, the officials of Mt. H0pe and Evergreen Cemeteries, and many Lansing and East Lansing residents who helped me find robin nests or let me enter their yards. Many of the statistical calculations were done using the Hang 600-14 calculating machine located in the Michigan State University Museum. I also wish to thank the graduate students in the Michigan State University Museum for many helpful comments and suggestions. A great deal of credit must go the members of my guidance committee, Dr. Donald J. Hall, Dr. Walter H. Conley, Dr. Donald L. Beaver, and especially my major professor, Dr. Richard H. Hill, for guidance of this thesis project and for help in preparing this manuscript. Finally, I wish to thank my wife Gayle for helping me when I needed an extra hand and for her tolerance of the many early awakenings. ii TABLE OF CONTENTS Eagg_ List of Tables .......................... iv List of Figures ......................... v Introduction ........................... 1 Methods and Materials ...................... 4 Results ............................. 17 Discussion ............................ 29 Suggestions for Future Research ................. 39 Literature Cited ......................... 40 iii Table Table Table Table Table Table Table Table Table 45mm LIST OF TABLES Characteristics of study plots, 1974 .......... Characteristics of study plots, 1975 .......... Nest density and reproductive output Number of nests fledging 0, l, 2, 3, relation to percentage lawn area for Nest density and reproductive output lS-acre plots in 1975 ................. Number of nests fledging 0, l, 2, 3, relation to percentage lawn area for in 1975 ........................ Nest density and reproductive output lO-acre subplots in 1975 ................ Number of nests fledging 0, l, 2, 3, relation to percentage lawn area for subplots in 1975 .................... Nest density and reproductive output combined results of lO-acre plots in subplots in 1975 .................... iv per plot for 1974 . and 4 young in l974 ....... per plot for and 4 young in lS-acre plots per plot for and 4 young in lO-acre per plot for 1974 and lO-acre $193 7 8 18 25 27 28 Figure 1. Figure 2. LIST OF FIGURES Aerial photograph of plot lA Aerial photograph of plot 7A INTRODUCTION Many aSpects of the biology of the American robin (Turdus migra- tgriu§_L.) have received attention, including breeding behavior, demo- graphics, migratory movements, and the effects of pesticides. Howell (1942) did an extensive study of the breeding behavior and nesting of the robin. Farner (1945a, 1945b, 1949) made use of extensive banding records to determine mortality rates, longevity, and tendency of robins to return to their birthplace. Speirs (1946, 1953) examined the local and migratory movements of the species. Young (1951, 1955) carried out detailed studies of territoriality and did further work concerning their breeding and nesting behavior. Mehner (1958) repeated many of the life history studies, but unlike the earlier researchers examined robins in residential areas. The effect of the insecticide DDT on robins has been examined, and the results have provided further evidence of the lethal effects of DDT in wildlife (Wallace, Nickell, and Bernard, 1961). James and Shugart (1974) examined robin nesting records in order to analyze climatic conditions necessary for initiation of nesting. In addition, the robin has been the subject of numerous short papers and anecdotal notes covering a time span of at least a hundred years. Despite the abundance of information concerning this Species, a great deal has yet to be learned about why robins occur where they do and the factors which tend to limit their p0pu1ation size. Barrows (1912) 2 wrote that it is a matter of common remark that they become much more abundant with the clearing and settling of a region by man. Yet the species is well known in the most densely forested portions of Michigan and in open, unsettled grounds. Howell (1942) noted that deSpite the robin's toleration of a wide variety of habitats, the changes which man introduces seem usually to favor it. In suburban areas, open lawns are very productive feeding grounds, and shade trees provide successful nest sites. Nickell (1944) wrote that the greatest concentrations of nesting pairs usually are found in cemeteries, orchards, parks, camps, estates, lake settlements, and towns. Ddum and Burleigh (1946) stated that although the breeding range of the robin in North America was probably fairly restricted at the time Columbus discovered America, the species has followed the development of man-made lawns, towns, and farmlands, increasing its distribution greatly. They suggested that this change in distribution provides good evidence that the vegetational habitat was a critical limiting factor in the original range of the robin. In 1958, Mehner confirmed the greater abundance of robins in residential areas as compared to natural habitats. Graber and Graber (1963), in a report on the population changes over a fifty-year period in Illinois, found that between 1907 and 1909 robin densities in residential areas averaged 32.5 individuals per 100 acres, whereas between 1957 and 1958, after increased urbanization, robin densities in residential areas averaged 114.2 individuals per 100 acres. The authors suggested that as farming declined in residential areas, mowed lawns and tree-lined streets gradually replaced grazed pastureland in providing arboreal nesting sites and clipped-grass foraging areas and, in fact, were much superior to pastureland in respect to these important resources. 3 In summary, it appears that man has played an important role in increasing the carrying capacity of the environment for robins. The exact reason for this, although there are several clues, is not known. Since robins feed almost exclusively on the ground, one possible explanation is that in the process of clearing the land for his own use, man has increased the amount of available food which adult robins can collect, for both themselves and their young. The purpose of this study was to investigate this possibility, and others, by comparing the percentage of cleared area, or lawn, in residential locations with several parameters of robin populations. Lawn was defined as that area not occupied by buildings, streets, sidewalks, etc. If indeed lawn area is a limiting factor, it might be expected that if all other factors are equal, the density and perhaps also the reproductive output of robins per unit of total area would depend upon the percentage of the total covered by lawn. Specifically, I hypothesized that (l) the density and (2) the reproductive success of breeding pairs of robins would decrease as the percentage of lawn area decreased. METHODS AND MATERIALS Field studies were conducted during the spring and summer of 1974 and 1975. All study sites were located in Lansing and East Lansing, Ingham County, Michigan. Nine 10-acre study plots were selected for study during 1974. They included a range of lawn areas from 44 percent to 100 percent. They were chosen by examining aerial photographs which were obtained from the city planners of Lansing and East Lansing. These photographs had been taken by local aerial photo services in 1968 and 1970. Seven lS-acre study plots independent of those in 1974 were selected for study during 1975. They included a range of lawn areas from 46 percent to 100 percent and were selected in the same manner as the plots for 1974. The rationale for the increase in study plot size in 1975 as compared to 1974 will be discussed presently. PrOSpective study plots were selected initially from aerial photo- graphs on the basis of size, shape, and percentage of lawn area. Recognizing that robin densities range from 0.57 to 1.2 pairs per acre in residential areas (Mehner, 1958; Graber and Graber, 1963), plot size was standardized at 10 acres in 1974. This size would theoretically provide between 18 and 37 attempted nests per study plot over the spring and summer based on published nesting rates (Young, 1955). This would be a satisfactory number of nestings for making statistical compari- sons. For consideration, a prospective study plot had to be of the 4 5 ten-acre size and display a relatively uniform distribution of lawn area throughout. The plots were square to rectuangular in shape and, except for the study plot within the golf course, were bounded by paved streets. Residential plots contained between two and four city blocks. Because of a lower-than-expected number of nestings in 1974, study plots were enlarged to 15 acres in 1975. Residential plots for this year contained between two and six city blocks, but were otherwise comparable to the 1974 study plots. In order to assure that the results from the two years could be combined and compared, each lS-acre study plot of 1975 had a lD-acre subplot blocked off, a priori, which could be compared statistically to the study plots of 1974. These lO-acre subplots included a range of lawn areas from 45 percent to 100 percent. ProSpective study plots of both years were examined in person to assure that no large changes had taken place since the time of the aerial photographs. Plots which contained large hills or slopes were rejected. In addition, the approximate number of trees on each plot was determined to assure that all plots would contain similar numbers of potential nest sites. Plots on which land development had occurred in recent years were considered unsatisfactory because these plots contained relatively few older and larger trees, which robins frequently use for nest sites. The types of trees varied considerably among study plots, but there is considerable evidence that robins select a variety of tree types for nesting sites and thus do not appear to be restricted by tree type (Howell, 1942; Nickell, 1944). Finally, any potential study plots which included major traffic routes were rejected in order to keep automobile activity levels as similar as possible among plots. 6 Once a study plot had been chosen, its exact lawn area was calcu- lated. This was done by copying the aerial photograph of the study plot on paper of uniform weight per unit area using a Minolta series 200 copying machine, separating the lawn from the nonlawn area with a razor blade, and then weighing the two. Throughout these procedures the sum of the lawn and nonlawn weights was compared to original total weight to be certain of accuracy. This procedure was followed for all plots except those of 100 percent lawn area. In the latter cases it was known from first-hand observation that negligible nonlawn area existed. Percent lawn areas of the study plots of 1974 and other information gathered during the plot selection period for plots of that year are presented in Table 1. Similar information for the study plots of 1975 is presented in Table 2. Hereafter, the nine study plots of 1974 will be referred to as plots 1A through 9A, whereas the seven 15-acre study plots of 1975 will be referred to as plots 18 through 78. The seven lO-acre subplots of 1975 will be referred to as plots 1C through 7C. Aerial photographs of plots 1A (44 percent lawn area) and 7A (73 percent lawn area) are shown in Figures 1 and 2, respectively, and illustrate the magnitude of difference between residential habitats of different percentage lawn area. Because of the nature of the Lansing-East Lansing area, it was impossible to select plots which were completely equivalent for all factors other than percentage lawn area. A lack of equivalence arose from several causes. First, study plots were chosen from a variety of residential locations, all having developed in at least slightly different manners and certainly without thought of experimental design with respect Table 1. Characteristics of study plots, 1974 Approximate Plot Percent number number lawn Description of trees 1A 44 residential 200 2A 58 residential 230 3A 58 residential 245 4A 59 residential 260 5A 66 residential 210 6A 70 residential 200 7A 73 residential 270 8A 88 cemetery 190 9A 100 golf course 150 Table 2. Characteristics of study plots, 1975a Approximate Plot Percent number number lawn Description of trees lB(lC) 46(45) residential 320(210) 28(2C) 49(48) residential 370(260) 3B(3C) 51(50) residential 300(230) 4B(4C) 67(66) residential 320(200) 53(5C) 7l(73) residential 340(230) 63(6C) 90(89) cemetery 290(200) 7B(7C) 100(100) golf course 200(170) aData in parenthesis are for 10-acre subplots. Figure 1. Aerial photograph of plot lA Figure 1 L -n‘ an..." ” V Figure 2. Aerial photograph of plot 7A 12 to robins. Thus, there were irreconcilable differences among study plots because of such factors as the shape and size of houses, the socio-economic backgrounds of the residents, and the environments surrounding the study plots. Second, unavoidable correlations between lawn area and other factors presented potential problems. As the percentage of nonlawn area increased, for example, human activity tended to increase. Third, in order to include study plots with a very high percentage of lawn area, it was necessary to select plOts in cemeteries and a golf course. Once again, equivalence among study plots was some- what compromised. Plots in cemeteries and the golf course contained fewer trees per unit area than residential plots. In spite of these difficulties, every attempt was made to maintain similarity among plots 'for factors other than percentage lawn area. It is hoped that the effects of all factors which are not equivalent among study plots will be offset, at least partially, by replication effects created by study plots of similar, but not identical, percentage lawn areas which were included in the two years of data collection. Prior to collection of data, circulars were distributed to houses within the study plots. The purpose of these circulars was to request the aid of residents in locating robins and robin nests and also to inform the residents that I would be in their neighborhoods throughout the spring and summer, usually equipped with binoculars and usually early in the morning. Since the working hypothesis was that food is a limiting factor for breeding success, the parameters selected for study were those that seemed most likely to reflect adequacy of food.) Data were gathered only on those breeding pairs which established a nest with eggs, 13 thereby inferring that they had sufficient resources for both maintenance and reproduction. The densities of such pairs and their reproductive output were compared among plots in the belief that these factors should reflect differential levels of resources, again, for both maintenance and reproduction. Specifically, the data collected were (1) attempted nests per plot, here defined as the number of nests in each plot which reached at least the egg stage, (2) successful nests per plot, or the number of nests in each plot which fledged at least one young, and (3) the number of young fledged per nest and, by extension, per plot. A parameter calculated from these data was (4) the average number of young fledged fbr each successful nest in each plot. The average number of young fledged per attempted nest was not examined because it would be partly a function of complete nesting failures, and complete failures are often catastrOphic and therefore not necessarily related to available food. Any relationship between nesting failures and percentage lawn area could be examined separately if desired. In addition, observations were collected on (1) nest location, (2) individual nest progress, (3) robin behavior, and (4) certain other parameters such as foraging habits and nest distribution which might help to explain the final results. Since birds were not banded, individuals could not be distin- guished during behavioral observations. Male and female robins could usually be distinguished from one another by the male's darker color. Young robins could be distinguished from adults by their Speckled breasts. Observations began during the first week of May, 1974, and during the last week of April, 1975, times that coincided with the initiation 14 of nest building by the birds. Thereafter, observations were made on each plot every three days insofar as weather and scheduling would permit and often more frequently when data had to be collected on an active nest. Observation periods were usually in the morning between 5:00 A.M. and 9:00 A.M., when the robins were most active, and usually lasted from 30 to 60 minutes per study plot. Observations were also frequently made in the evening from 6:00 P.M. to approximately 9:00 P.M. Since the leaves in early May were for the most part still at the bud stage, the first nests were found quite easily. Throughout the data collection period nests were located by watching for robins that were building nests, listening for males calling, talking to residents of the neighborhoods, and watching robin activity in general. Many robins could be observed foraging and then followed to the nest site. Nest site locations were recorded on maps of the study areas. Once a nest was located, its condition and contents were determined. This was usually done by using a collapsible 12-foot aluminum pole with a small hand mirror taped to one end. When necessary, a 16-foot wooden extension was lashed to this pole to give added height. The image of the nest contents (eggs or nestlings) was viewed in the mirror by using binoculars. An alternative method of monitoring nest contents, but one which could only be used once the eggs had hatched, consisted of counting the heads of the young as they were being fed. When the parents landed on the nest to feed the young, the nestlings extended their heads upward to receive food and at that point were usually visible with the aid of binoculars. Although this method was possibly not as reliable as observing the nestlings with the mirror, observations of this type were usually consistent for any one nest and on several 15 occasions were confirmed by using the mirror or another more direct technique. Occasionally the tree in which a nest was located was climbed if the contents could not be ascertained in any other way. The number of fledglings per nest was determined by following the progress of each nest and recording the number of nestlings last seen before fledging. The time of actual fledging was not completely predictable, and fledging was rarely observed. Nestling counts on nests near fledging were made each day. Regardless of the method used to enumerate nestlings, several pre-fledging counts were made fbr each nest as it was easy to overlook one or even two youngrobins as they sat in the crowded nest. The number of fledglings of six nests (four in 1974 and two in 1975) could not be determined by the usual procedure due to problems of accessibility (height, foliage density, and presence of power lines). Fledge counts for these nests were estimated from (1) nestling counts, which were likely to be inaccurate because of the inaccessibility, and (2) the number of fledglings subsequently seen off the nest being fed. Young robins usually remain around the nesting area with their parents for several days after fledging, and it proved possible to gain an estimated fledge count by these methods for all six nests for which accurate data could not be obtained in any other manner. These esti- mates are identified in the results where they appear. Throughout the study the nesting birds were disturbed as little as possible. The reaction of parent robins to my observations varied greatly among individuals. In some instances I was attacked by the parent robins, while at other times the robins simply observed the observer. 16 Observations were terminated in the second week of August in both 1974 and 1975, after nearly two weeks during which no new nests had appeared. Howell (1942), Young (1955), and Mehner (1958) all gave evidence that the breeding season of robins normally extends only until late July. RESULTS The results will be presented and analyzed statistically in the following order: (1) data from 1974, (2) data from lS-acre plots of 1975, (3) data from lD-acre subplots of 1975, and (4) combined data of 1974 and 1975 using the lO-acre subplots of 1975. Data on attempted nests, successful nests, total young fledged, and young fledged per successful nest for each plot were examined statistically. The hypothesized relationship between each of these parameters and percentage lawn area was tested using linear regression analysis [Y = a + b(SE)X], where Y is the dependent variable, X is the independent variable, a is the Y-intercept, and b is the regression coefficient. The F-statistic (Sokal and Rohlf, 1969) was used to determine to what extent, if any, the variances of these parameters were explained by regression on percentage lawn area. The null hypo- thesis for this test was: b = 0. The alternative hypothesis was: b > 0. The null hypothesis was rejected if P < 0.05. The appropriate regression equation is included in the results whenever a significant relationship is present. In 1974, a total of 52 attempted nests (nests with eggs) was observed within the nine study plots. Of these, 42 nests, or 81 percent, were successful, fledging a total of 94 young. Thus, 2.2 young were produced for every successful nest. Individual plot data are presented in Table 3. l7 18 Table 3. Nest density and reproductive output per plot for 1974a Average number Number of Number of Number of of young Plot Percent attempted successful total young fledged per number lawn nests nests fledged successful nest 1A 44 2 l 3 3.0 2A 58 5 4 10 2.5 3A 58 4 3 8b 2.7 4A 59 7 6 12C 2.0 5A 66 6 6 12 2.0 6A 70 6 6 10" 1.7 7A 73 8 5 10 2.0 BA 88 7 4 14 3.5 9A 100 7 7 15 2.1 Totals 52 42 94 2.2 aIndividual nest data presented in Table 4. bIncludes one estimated fledge count. CIncludes two estimated fledge counts. 19 The number of nests attempted per plot decreased as the percentage of lawn area decreased. The regression equation is Y = 0.45 + 0.078(20.0066)X. Analysis of this relationship using the F-statistic shows that a significant (P < 0.025) portion of the variance of attempted nests is explained by regression on percentage lawn area. Much of this significance seems to be due to low values for plots 1A, 2A, and 3A as shown by Table 3. The number of successful nests per plot also decreased signifi- cantly with percentage lawn area (P < 0.05). The regression equation is Y = -0.14 + 0.07(:0.0073)X. As seen in Table 3, this relationship is again due mostly to low values on plots 1A, 2A, and 3A. The total number of young fledged per plot also decreased as percentage lawn area decreased. Herein lies the most significant result of the 1974 data (P < 0.005). The regression equation is Y = -l.52 + 0.17(t0.0099)x. This result is most impressive because, as shown in Table 3, the relationship is produced not only by low numbers of young fledged at low percentage lawn areas but also by intermediate numbers of young fledged at intermediate percentage lawn areas and by high numbers of young fledged in locations of high lawn percentage. Finally, the number of young fledged per individual successful nest was analyzed in relation to percentage lawn area. The data are contained in Table 4. The number of young fledged per successful nest did not vary significantly with percentage lawn area (P > 0.75). Average values for young fledged per successful nest contained in Table 3 reflect this conclusion. Although there are insufficient data to make statistical comparisons, the number of complete nesting 20 Table 4. Number of nests fledging 0, l, 2, 3, and 4 young in relation to percentage lawn area for 1974 Plot Percent number lawn 0 young 1 young 2 young 3 young 4 young 1A 44 1 0 O l 0 2A 58 l 1 l 1 1 3A 58 l 0 2 0 1 4A 59 l 2 2 2 0 5A 66 0 2 2 2 0 6A 70 0 3 2 1 0 7A 73 3 2 l 2 0 8A 88 3 0 0 2 2 9A 100 0 3 l 2 1 Totals 10 13 ll 13 5 21 failures also appear to show no relationship to percent lawn area (see Table 4). In 1975, a total of 83 attempted nests was observed within the seven study plots. Of these, 68 nests, or 82 percent, were successful, fledging a total of 214 young. Thus, 3.1 young were produced for every successful nest. Individual plot data for 1975 are presented in Table 5. Unlike in 1974, the number of nests attempted per plot did not decrease significantly as the percentage of lawn area decreased (P > 0.75). This conclusion is also evident in simple inspection of Table 5. Plot 18 (46 percent lawn area) had one of the highest values for attempted nests while plot 7B (100 percent lawn area) had one of the lowest values for this same parameter. The number of successful nests per plot also did not decrease significantly with percentage lawn area (P > 0.75). To a great extent these data reflect the results on attempted nests (see Table 5). Once again, these results differ from those obtained in 1974. The total number of young fledged per plot also did not decrease significantly as percentage lawn area decreased (P > 0.75). Table 5 shows that nests from the two study plots of lowest percentage lawn area (46 percent and 49 percent) fledged more total young than nests from the two study plots of greatest percentage lawn area (90 percent and 100 percent). The number of young fledged per individual successful nest is shown in Table 6. The results are not significantly explained by regression on percentage lawn area (P > 0.75). This conclusion is reflected in the data on average young fledged per successful nest found in Table 5. Like the data of 1974, these results indicate that 22 Table 5. Nest density and reproductive output per plot for 15-acre plots in 1975a Average number Number of Number of Number of of young Plot Percent attempted successful total young fledged per number lawn nests nests fledged successful nest 13 46 14 11 36b 3.3 28 49 12 10 33 3.3 38 51 8 7 23 3.3 48 67 14 10 29b 2.9 58 71 16 13 36 2.8 68 90 ll 9 27 3.0 78 100 8 8 28 3.5 Totals 83 68 214 3.1 aIndividual nest data presented in Table 6. bIncludes one estimated fledge count. 23 Table 6. Number of nests fledging O, l, 2, 3, and 4 young in relation to percentage lawn area for lS-acre plots in 1975 Plot Percent Number lawn 0 young 1 young 2 young 3 young 4 young 1B 46 3 0 l 6 4 28 49 2 0 l 4 38 51 ‘ 1 o o 5 2 4B 67 4 l 1 6 2 5B 71 3 0 4 8 l 68 90 2 0 1 7 1 7B 100 O 0 O 4 4 Totals 15 l 8 41 18 24 the number of young fledged per nest is not a function of percentage lawn area. The number of complete nesting failures per plot as shown in Table 6 is in agreement with this result. It should be noted that, on the average, reproductive success was much higher in the study plots of 1975 than in those of 1974 (compare Table 4 and Table 6). In 1975 a greater percentage of nests fledged 3 and 4 young than in 1974. The results from the individual lO-acre subplots of 1975 are contained in Table 7. These results are reviewed in detail here because they do not, in all instances, simply reflect the data from the lS-acre plots. The number of nests attempted per plot did not decrease as the percentage lawn area decreased, the F-statistic for the regression being nonsignificant (P > 0.75). This result disagrees with the results of 1974 and agrees with the results from the lS-acre study plots of 1975. Note that the number of attempted nests recorded in plot 3C, and also 6C, was disprOportionately low, given the numbers of nests in plots 38 and 6B (compare Table 5 and Table 7). Possible explanations for this phenomenon and the implications it holds for this study will be discussed later. The number of successful nests per plot also did not decrease significantly as the percentage of lawn area decreased (P > 0.75). This result disagrees with the 1974 results and agrees with the results from the lS-acre study plots of 1975. As is shown in Table 7, the number of successful nests is for the most part, unifonm among study plots. The relatively low numbers of successful nests in plots 3C and 6C reflect the low numbers of attempted nests in these plots. 25 Table 7. Nest density and reproductive output per plot for lO-acre subplots in 1975a Average number Number of Number of Number of of young Plot Percent attempted successful total young fledged per number lawn nests nests fledged successful nest 1c 45 11 9 30b 3.3 2C 48 9 8 30 3.8 3C 50 2 2 6 3.0 4C 66 12 8 23 2.9 SC 73 12 10 28 2.8 6C 89 5 5 16 3.2 7C 100 7 7 24 3.4 Totals 58 49 157 3.2 aIndividual nest data presented in Table 8. bIncludes one estimated fledge count. 26 The number of young fledged per plot also did not decrease signifi- cantly as the percentage of lawn area decreased (P > 0.75). This result disagrees with the analogous result of 1974 and agrees with the result fbr 15-acre plots in 1975. As shown in Table 7, the results are once again relatively uniform among study plots except for plots 3C and 66. The numbers of young fledged per individual successful nest are shown in Table 8. As with the results from both 1974 and the 15-acre plots of 1975, the results are not significantly explained by regression on percentage lawn area (P > 0.75). This result is reflected in the data on average young fledged per successful nest contained in Table 7. The data on complete nesting failures found in Table 8 also do nothing to contradict this result. Lastly, the combined results from the lO-acre plots of 1974 and the lO-acre subplots of 1975 will be briefly presented. Data on attempted nests, successful nests, total young fledged, and young fledged per successful nest for each plot for these two years are found in Table 9. The hypothesized relationship between each of these parameters and percentage lawn area was found in each case to be nonsignificant (P > 0.75). 27 Table 8. Number of nests fledging 0, l, 2, 3, and 4 young in relation to percentage lawn area for lO-acre subplots in 1975 Plot Percent number lawn 0 young 1 young 2 young 3 young 4 young 1C 45 2 0 l 4 , 4 2C 48 1 0 l 3 3C 50 0 0 0 2 0 4C 66 4 1 1 4 2 5C 73 2 0 3 6 1 6C 89 0 0 0 4 1 7C 100 0 0 O 4 3 Totals 9 l 6 27 15 28 Table 9. Nest density and reproductive output per plot for combined results of lO-acre plots in 1974 and 10-acre subplots in 1975 Average number Number of Number of Number of of young Plot Percent attempted successful total young fledged per number lawn nests nests fledged successful nest 1A 44 2 1 3 3.0 16 45 11 9 3oal 3.3 2C 48 9 8 30 3.8 3C 50 2 2 6 3.0 2A 58 5 4 10 2.5 3A 58 4 3 8“ 2.7 4A 59 7 6 12b 2.0 5A 66 6 6 12 2.0 4C 66 12 8 23 2.9 6A 70 6 6 10"“ 1.7 7A 73 8 5 10 2.0 SE 73 12 10 28 2.8 8A 88 7 14 3.5 6C 89 5 5 16 3.2 9A 100 7 7 15 2.1 7C 100 7 7 24 3.4 Totals 110 91 251 2.8 aIncludes one estimated fledge count bIncludes two estimated fledge counts DISCUSSION Prior to discussion of the immediate implications of the results obtained in this study, the densities and reproductive success observed should be compared to those seen in other studies of robins. Throughout these comparisons, the results of other researchers have been manipulated arithmetically, where necessary, to comply with the definitions of the present study. In general, fewer robin nests were observed in 1974 than had been expected based on reported research in residential areas (see page 4). Roughly, the robin densities in 1974 stood between 0.3 and 0.4 robin pairs per acre in study plots of greatest robin abundance. Low density is the reason why plot size was increased to 15 acres per study plot in 1975. The nest densities of 1975 were not any greater than some of those observed in 1974. There are several possible explanations for the lower-than-expected densities. First of all, cold and rainy weather early in the Spring of 1974 apparently affected nesting success and may have also affected nest density. Poor weather was probably respon- sible for the lower average number of young fledged per successful nest in 1974 as compared to 1975. Mehner (1958) noted lower success due to weather, and Howell (1942) noted that, on the average, the first brood of a pair of robins is less successful than later broods because of uncertain weather conditions in the early Spring. The effect of weather on nesting densities, if any, is not clear. Young (1955) wrote 29 30 that cold and rainy weather is probably a factor during most breeding seasons. Yet, there are no published reports of low densities early in the spring. If weather was preventing robins from nesting early in 1974, then robin densities should have increased as the breeding season progressed. Such an increase, however, was not observed. In 1975 the Spring was warm and dry, yet densities remained comparable to 1974. Thus, it seems unlikely that weather conditions provide a viable explana- tion for the lower-than-expected densities. Secondly, and more likely, it is possible that the robin p0pulations in the Lansing-East Lansing area are still suffering from the losses caused by spraying of DDT to control Dutch elm disease. During the middle and late 1950's robin populations were completely eliminated in certain areas of Lansing and East Lansing (Wallace, Nickell, and Bernard, 1961). While these birds are no longer receiving lethal doses of this poison, it is possible that the populations have yet to return to previous levels. Unfbrtunately, this problem cannot be examined using the results from this study. A final point worthy of note is that it is always difficult to predict breeding bird densities for a particular location based on densities from another time and location. With respect to carrying capacity, the situation existing in the Lansing-East Lansing area may not be analogous to other situations where data have been collected on robins. It should be emphasized here that for all statistical comparisons, density was measured only indirectly using the number of attempted nests. Since nesting efforts do not always begin at the same time, it is difficult to infer actual density using nest counts. Unless, however, there are large numbers of nesting failures, and therefore renesting by identical 31 pairs, the number of attempted nests provides a good indicator of relative densities. The 81 percent rate of nesting success found in 1974 and the 82 percent rate found in 1975 are higher than occur in most other passerine Species. The results of other researchers working with robins have varied greatly. Kendeigh (1942) reported 78 percent successful nests while Howell (1942) found a success rate of 61 percent. Young (1955) reported an average success rate of 49 percent, and Mehner (1958) found success rates ranging from 48 percent to 100 percent in different residential locations. Success rates found in this study are within the range of values found by other researchers but are located near the high end of this range. The fledging rate of robins has also been recorded by a number of researchers, and again, the results vary greatly. Howell (1942) found that, on the average, 1.7 young were fledged for every successful nest. Young (1955) and Mehner (1958) reported values of 2.9 and 2.4, respectively. The results of this study are similar to these data (2.2 in 1974 and 3.1 in 1975). In summary, although the density of robin nests observed in this study was lower than expected, the nests which were observed showed reproductive success comparable to that found in other robin p0pu1ations as reported by other researchers. It is thus hoped that the discussion of the results will be applicable to robins of various other times and locations. The discussion of the results must be in terms of (l) proximal factors, those factors which birds themselves use to select a habitat, and (2) ultimate factors, those factors which are related to the bird's 32 requirements and which eventually determine whether or not the species can survive and reproduce in a particular habitat (Hilden, 1965). It is fairly well established that, proximally, robins tend to select a nest Site near their birthplace or near their nest site of the previous year (Hickey, 1943; Farner, 1945b). In addition to research directed specifically at this phenomenon, reports and notes regarding the phenomenon are too numerous to mention. The tendency of robins to return to previous sites is important in this study because of its implications for both expected and observed results. Actual knowledge about the factors which robins require ultimately to survive is not so abundant. The purpose of this study was to test whether or not lawn area, or some correlate of lawn area, is a limiting factor during the breeding season. Based on this hypothesis and know- ledge about the way robins select habitat proximally, one or more of the following results would be expected. (1) A decrease in the number of attempted nests with a decrease in percentage lawn area. Since robins tend to return to the location of either their birth or previous nesting, it would be expected that, on the average, fewer robins should attempt nests in locations of low percentage lawn because of relatively low levels of survivorship and/or reproductive success in preceding years due to the hypothesized limiting factor. In instances where robin densities are high relative to land area, one has to assume in predicting higher densities with greater percent lawn that robin territories are not of evolutionary predetermined Size and that the birds are thus able to "pack in" at locations of high percentage lawn area. Certain birds have compressible territories while others do not (Helty, 1962). Although robin territory size does vary with location (Howell, 1942; 33 Young, 1955), it is not known if territories within any one location are compressible. (2) A decrease in the number of successful nests with a decrease in percentage lawn area. If the number of attempted nests decreases with percentage lawn area owing to the factors just discussed, the number of successful nests might be expected to decrease in tandem. Even if the number of attempted nests does not decrease, however, the number of successful nests would decrease with percentage lawn area due to a lower percentage of success with each accompanying decrease in percentage lawn area. (3) A decrease in the total number of young fledged with a decrease in percentage lawn area. This could result from a decrease in the number of attempted nestings, decrease in the number of successful nestings, and/or a decrease in the number of young fledged per successful nest, again because of some limiting factor associated with percentage lawn area. (4) A decrease in the average number of young fledged per successful nest with a decrease in percentage lawn area. If there is some limiting factor related to percentage lawn area, parent robins should on the average, bring fewer healthy nestlings per brood to the fledging stage in locations of low percentage lawn area. Comparison of the expected results with the observed results shows that the 1974 data support the hypothesis that lawn, or some correlate of lawn, is a limiting factor for breeding robins. The numbers of attempted nests, successful nests, and total young fledged all decreased with decreases in percentage lawn area. The 1975 results which include larger study plots and higher numbers of total attempted nests, do not provide support for the hypothesis that percentage lawn is related to a limiting factor for breeding robins. The data from the 10-acre 34 subplots of 1975 and the combined results of 1974 and 1975 both reinforce this conclusion. The rationale for combining the results from the two years was to test the initial hypotheses using a large number of study plots of which no two were exactly alike. Recognizing that there are indications that data from the two years differ because of factors in addition to time, namely weather, discussion and conclusions based on the combined results will be held to a minimum. In view of the individual 1974 and 1975 results the original hypothesis is rejected. Percentage lawn area is a necessary but not a sufficient factor in any model attempting to predict robin density and reproductive output. In the remainder of the discussion an attempt will be made to reconcile the contradictory results of 1974 and 1975. It appears from the data that if lawn, or some correlate of lawn, is a limiting factor for breeding, it is not a limiting factor during all years. It has often been suggested, however, that limiting factors need not be, and indeed are not, limiting in all years. Thus, p0pu1ations are only limited in certain years when density independent factors, such as weather, are particularly harsh, or when density dependent factors, such as food, are not abundant relative to p0pu1ation numbers. Such a process could be one possible explanation for the observed results. Thus, 1974 data may have been gathered in a year when some limiting factor was operating, whereas this same factor may not have been limiting in 1975, at least not at percentage lawn areas greater than 46 percent. Since robin nesting density and success decreased with a decrease in percentage lawn area in 1974, any proposed limiting factor would have to be one which explained these phenomena. One possibility is that available food, affected by weather, was limiting in 1974. It has already been noted 35 that the weather was unfavorable during the spring of 1974, and that this weather was probably responsible for the lower average number of young fledged per successful nest in 1974 as compared to 1975. It has also been noted that robins feed primarily on the ground. Cold, damp weather during the spring could affect the level at which soil invertebrates are found and, in addition, affect emergence rates of many other invertebrates, including insects. Howell (1942) reports that approximately 70 percent of the diet of robins is composed of invertebrates. A greater amount of lawn area in close proximity to the nest might, on the average, allow parent robins to successfully raise more young per brood. Heather, as a limiting factor in itself, does not sufficiently explain why robin nesting density and reproductivity increased as percent lawn area in- creased in 1974 because it would be expected with a limiting factor of this type that all plots would Show similar effects. Thus, it is proposed that weather was acting only indirectly. Hypothesizing food as a limiting resource in 1974 does not in itself explain the observed decrease in the number of attempted nests with a decrease in percentage lawn area in the 1974 results. One explanation is that 1973 was also a food-limited year and that robin pairs in locations of low percentage lawn area in that year were, on the average, less successful than pairs in locations of high percentage lawn area. Thus, there were fewer robins to return to locations of low percentage lawn area as compared to numbers returning to locations of high lawn percen— tage. If this were true, this same chain of events Should have occurred during 1974 and 1975; yet, the number of attempted nests did not decrease with a decrease in percentage lawn area in 1975. Another explanation for the observed decrease in the number of attempted nests with a decrease in 36 percentage lawn area in 1974 is that robins are somehow able to measure proximally the abundance of available food before they nest. Thus, in addition to returning to the site of the previous year, they might also adjust for seasonal conditions. While there are data which indicate that some birds, specifically ovenbirds, might be capable of gauging the abun- dance of invertebrates at the time of nesting and acting apprOpriately (Stenger, 1958), there is no evidence that robins act similarly. A second explanation for the observed results of 1974 and 1975 is that percent lawn area is not associated with a limiting factor for breeding. Thus, the results of 1974 were a chance occurrence and the results of 1975 more accurately portray the biology of this situation. While chance occurrence is an alternative in any experiment using statistics, I wish to point out that the fact that robins are very locally oriented may have some particular bearing on the idea that the 1974 results were a chance occurrence. AS was previously mentioned, robins tend to return to the site of their bird or previous nesting. In addition, the one to three broods which robins raise during the Spring and summer are often within 100 feet of one another (Howell, 1942). This lack of dynamic movement could create breaks in local distribution when birds die. Thus, if a robin or robin pair is killed in migration or during the breeding season, certain locations may lack nesting robins, for all or a portion of a breeding season. Other locations which do not suffer similar sets of random circumstances may be very productive in terms of robins because of the tendency of robins to bunch the nests of a single season together. Thus, any chance occurrence of this type becomes magnified in the results when measured as they were here. Because of this problem, the number of 37 attempted nests and their reproductive success for any one location may be historically linked to an especial extent and may not reflect the conditions dictated by an hypothesized limiting factor. This phenomenon, which would not be expected to affect results based on a large number of nests, may explain the 1974 results. An example of this possibility is found in the results of 1975. AS was noted previously, the results from plots 3C and 6C do not give a true indication of the nesting densities of the larger, 15-acre plots, 3B and 68, respectively. I believe that this is due to the local affinities of robin nesting pairs. As can also be seen, the data for plots 3C and 6C are not unlike some of the data from 1974. In spite of this, however, this phenomenon is unlikely to explain trends such as those seen in 1974. Whether the results found here can be applied to robins of other times and locations remains questionable. It would appear from the research done both here and elsewhere that robin densities are highly variable. Results and observations presented here indicate that robins, in at least some locations, are somewhat sparingly distributed. At no time did the robins in this study approach densities which have been fOund by previous workers. Some of the data of Young (1955) indicate vacant areas between existing robin territories. Conversely, Wallace (1960, 1962) found that as robins died from DDT poisoning, other robins entered the area to occupy vacant territories. This would seem to indicate a floating p0pu1ation of non-breeding birds as found by Hensley and Cope (1951) for certain other species of birds. The presence of a floating p0pu1ation would seem to indicate an abundance of robins relative to some limiting breeding resource. Thus, not all birds are able to initiate breeding. Farner (1945b) felt that robins 38 do not have floating populations. If the birds in this study were limited by percentage lawn area, then we must wonder what factors allow robins to reach much higher densities in other locations. If the birds in this study were not limited by percentage lawn area, it may be because this limiting factor becomes Operative only at higher densities of birds. In light of the variability of robin densities, it is very difficult to draw conclusions about the factor or factors which are limiting this species. Indeed, it may vary. Thus, it is felt that the results found here, along with the accompanying discussion, should be applied to robins of other times and locations with some caution. Regardless of these considerations, however, it can be concluded that an increase in percentage lawn is not sufficient to assure an accom- panying increase in robin density and reproductive success. SUGGESTIONS FOR FUTURE RESEARCH Several possible suggestions can be made for further research on the t0pic. There seems to be some question whether robin density is reasonably predictable for any one location. Hence, it might prove fruitful to make observations on robin populations in a small number of study areas for a longer period of time than was done here. In this manner, a great deal more might be learned about the limiting factors of this bird. If robin numbers vary on the breeding range from year to year without apparent accompanying changes in the condition of the breeding area, it may be that robin numbers are regulated during the non-breeding season by°such things as the hardships of migration or food limitations on the wintering range. If robin numbers appear to remain relatively stable fbr any one location, then evidence about limiting factors might be obtained by analyzing habitats in which robins occur at a large range of densities during the breeding season. While this might include a great deal of travel, it may lead to immediate clues about what factor or factors are regulating robin numbers. The possibility that food is a limiting factor could be examined more directly by comparing food levels to density and reproductive output in robins. It may be that there is some threshold level of available food at which food becomes limiting for robins and that depending on weather, this threshold is associated with a variable percentage of lawn. 39 LITERATURE CITED LITERATURE CITED Barrows, W. B. 1912. Michigan Bird Life. Vol. 2. Michigan Agricul- tural College, East Lansing. (see pp. 722-727.) Farner, D. S. 1945a. Age groups and longevity in the American robin. Wilson Bull. 57: 56-74. Farner, D. S. 1945b. The return of robins to their birthplaces. Bird- Banding 16: 81-89. Farner, D. S. 1949. Age groups and longevity in the American robin: comments, further discussion, and certain revisions. Wilson Bull. 61: 68-81. Graber, R. R. and J. W. Graber. 1963. A comparative study of bird populations in Illinois, 1906-1909 and 1956-1958. Ill. Nat. Hist. Surv. Bull. 28: 485-486. Hensley, M. M. and J. B. COpe. 1951. Further data on removal and rep0pulation of the breeding birds in a spruce-fir forest community. Auk 68: 483-493. Hickey, J. J. 1943. A Guide to Bird Watching. Oxford University Press, New York. (see pp. 40-4T.) Hilden, O. 1965. Habitat selection of birds. Ann. 2001. Fenn. 2: 53-75. Howell, J. C. 1942. Notes on the nesting habits of the American robin (Turdus migratorius L.). Am. Midl. Nat. 28: 529-603. James, F. C. and H. H. Shugart, Jr. 1974. The phenology of the nesting season of the American robin (Turdus migratorius) in the United States. Condor 76: 159-168. Kendeigh, S. C. 1942. Analysis of losses in the nesting of birds. J. Wildl. Manage. 6: 24-42. Mehner, J. F. 1958. Studies on the life history of the robin (Turdus migratorius Linnaeus). Ph.D. Thesis, Michigan State University. East [ansing, Michigan. Nickell, W. P. 1944. Studies of habitats, locations, and structural materials of nests of the robin. Jack-Pine Warbler 22: 47-64. 40 4l Odum, E. P. and T. D. Burleigh. 1946. Southward invasion in Georgia. Auk 63: 388-401. Sokal, R. R. and F. J. Rohlf. 1969. Biometr . W. H. Freeman and Company, San Francisco. (see pp. 404-424.) Speirs, J. M. 1946. Local and migratory movements of the American robin in eastern North America. Ph.D. Thesis, University of Illinois. Urbana, Illinois. Speirs, J. M. 1953. Winter distribution of robins east of the Rocky Mountains. Wilson Bull. 65: 175-183. Stenger, J. 1958. Food habits and available food of ovenbirds in relation to territory size. Auk 75: 335-346. Wallace, G. J. 1960. Another year of robin losses on a university campus. Audubon Mag. 62: 66-69. Wallace, G. J. 1962. The seventh spring die-off of robins at East Lansing, Michigan. Jack-Pine Warbler 40: 26-32. Wallace, G. J., W. P. Nickell, and R. F. Bernard. 1961. Bird mortality in the Dutch elm disease control program in Michigan. Cranbrook Inst. Sci. Bull. 41: 1-44. Welty, J. C. 1962. The Life of Birds. W. B. Saunders Company, Phila- delphia. (see pp. 220-22T.T Young, H. 1951. Territorial behavior in the eastern robin. Proc. Linn. Soc. N.Y. Nos. 58-62: 1-37. Young, H. 1955. Breeding behavior and nesting of the eastern robin. Am. Midl. Nat. 53: 329-352. MICHIGAN STATE UNIV. LIBRARIES illlllllililllliiliiill)llU“Willillllllillllilill 31293006329043