BIOLOGY AND CONSERVATION OF THE COMMON MURRE 37 



regional population trends prior to 1979. To summarize 

 historical information, and without the aid of 

 computerized databases or search processes, we 

 consulted all available published and unpublished 

 sources known to us from prior research as follows: 

 California (H. R. Carter, see Appendix B), Oregon (R. 

 W. Lowe), Washington (U. W. Wilson and H. R. Carter), 

 and British Columbia (M. S. Rodway and H. R. Carter). 

 To augment historical information, H. R. Carter also 

 examined egg records in California, Oregon, 

 Washington, and British Columbia in major museum 

 collections, including Western Foundation of Vertebrate 

 Zoology, Camarillo, California (WFVZ); Humboldt State 

 University, Department of Wildlife Museum, Arcata, 

 California (HSUWM); Santa Barbara Natural History 

 Museum, Santa Barbara, California (SBNHM); San 

 Diego Natural History Museum, San Diego, California 

 (SDNHM); University of California Berkeley, Museum 

 of Vertebrate Zoology, Berkeley, California (BMVZ); 

 National Museum of Natural History, Smithsonian 

 Institution, Washington, D.C. (USNM); Museum of 

 Comparative Zoology, Harvard University, Cambridge. 

 Massachusetts (MCZ); Royal British Columbia 

 Museum, Victoria, British Columbia (RBCM); and 

 University of British Columbia Zoology Museum, 

 Vancouver. British Columbia (UBCZM). Substantial 

 historical information was rediscovered in these museum 

 collections. 



Annual survey data for each colony in 1979-95 

 were collated as follows: California (H. R. Carter and J. 

 E. Takekawa; see Appendices C, D), Oregon (R. W. Lowe; 

 see Appendix E). Washington (U. W. Wilson; see 

 Appendixes F and G), and British Columbia (M. S. 

 Rodway). Regional estimates of the total number of 

 murres attending colonies during surveys were 

 determined by summing single representative whole- 

 colony counts from aerial photographic surveys for each 

 colony within a particular year in central and northern 

 California and southern and northern Washington (see 

 Appendix A for survey methods and summary). For these 

 areas, we summed counts only when all or most colonies 

 were judged to have been surveyed in a generally 

 standardized and compatible fashion in the same year. 

 For Oregon, we summed counts for 15 sample colonies 

 that were surveyed and counted annually between 1988 

 and 1995. This sample of colonies is spread along the 

 entire Oregon coast, but several large colonies are not 

 included because of extensive counting time required. 

 Annual sums of whole-colony counts in each geographic 

 area are presented in Appendix H. In keeping with 

 seabird colony catalogs (Sowls et al. 1980; Speich and 

 Wahl 1989; Rodway 1 991: Carter etal. 1992). we referred 

 to specific colonies or subcolonies as they have been 



previously defined, which allowed easy cross referencing 

 between sources. 



We also summed whole-colony counts within 

 "colony complexes" in California and Washington. We 

 considered a colony complex to be a geographic subunit, 

 composed of several colonies close together. Such 

 subunits reflected major geographic assemblages of 

 breeding murres that resulted from the distribution of 

 suitable breeding habitat, accounted for the greater 

 potential for interaction between nearby colonies, and 

 accounted for inconsistent definitions of what 

 constituted a colony in seabird colony catalogs (Sowls 

 et al. 1980; Speich and Wahl 1989; Carter et al. 1992). 

 We lumped adjacent rocks, islands or mainland cliffs 

 with groups of breeding murres into colony complexes 

 when they were within about 5 km of each other. Colony 

 complex totals are presented in Appendix H. 



Regression analysis has been used extensively to 

 assess avian population trends and is a widely accepted 

 method of demonstrating and measuring the rate of 

 population change over a period of time (Sauer and 

 Droege 1990). We used regression analysis to calculate 

 rates of population change (percent per annum change 

 with 95% confidence intervals) for each area within 

 certain periods and to determine statistical significance 

 for population trends identified in these periods (see 

 Appendix H). We conducted linear regression analyses 

 on single, annual sums of whole-colony counts in 

 geographic areas over a period of years, only including 

 standardized and compatible data. Because of 

 availability of data and previously identified population 

 changes, we conducted regressions over all years of 

 available data between 1979 and 1995. as well as on 

 series of years where our direct inspection of data 

 indicated a distinct trend (i.e., increasing, decreasing, 

 or no change). This approach led to the following 

 regression periods: central California (1979-89. 1985- 

 95. 1979-95). northern California (1979-89), Oregon 

 (1988-95), and Washington (1979-86. 1984-95. and 

 1979-95). Significant regressions (P < 0.050) are 

 reported in the text and presented in figures. Where 

 significant trends were not detected, changes in 

 whole-colony counts are discussed in the text and 

 presented in figures. 



For certain objectives, we estimated the number of 

 breeding adults at a colony and summed colonies to 

 determine the size of a breeding population. 

 Whole-colony counts of murres can be adjusted with a 

 k correction factor to convert whole-colony counts to 

 either "number of breeding pairs" or "number of 

 breeding individuals" (Nettleship 1976; Birkhead and 

 Nettleship 1980; Sowls et al. 1980; Takekawa et al. 



