CALIFORNIA FTSFHGAME "CONSERVATION OF WILDLIFE THROUGH EDUCATION" VOLUME 73 JULY 1987 NUMBER 3 California Fish and Game is a journal devoted to the conservation and understanding of fish and wildlife. If its contents are reproduced elsewhere, the authors and the California Department of Fish and Game would appreciate being acknowledged. Subscriptions may be obtained at the rate of $10 per year by placing an order with the California Department of Fish and Game, 2201 Garden Road, Monte- rey, CA 93940. Money orders and checks should be made out to California Fish and Game. Inquiries regarding paid subscriptions should be directed to the Editor. Complimentary subscriptions are granted on an exchange basis. Please direct correspondence to: Robert N. Lea, Ph.D., Editor-in-Chief California Fish and Game 2201 Garden Road Monterey, CA 93940 u b VOLUME 73 JULY 1987 NUMBER 3 Published Quarterly by STATE OF CALIFORNIA THE RESOURCES AGENCY DEPARTMENT OF FISH AND GAME — LDA— STATE OF CALIFORNIA GEORGE DEUKMEJIAN, Governor THE RESOURCES AGENCY GORDON VAN VLECK, Secretary for Resources FISH AND GAME COMMISSION ALBERT C. TAUCHER, President Long Beach ABEL C. GALLETTI, Vice President JOHN A. MURDY III, Member Rancho Palos Verdes Newport Beach ROBERT A. BRYANT, Member E. M. McCRACKEN, JR., Member Yuba City Carmichael HAROLD C. CRIBBS Executive Secretary DEPARTMENT OF FISH AND GAME PETE BONTADELLI, Acting Director 1416 9th Street Sacramento 95814 CALIFORNIA FISH AND GAME Editorial Staff Editorial staff for this issue: Marine Resources Paul N. Reilly, Peter L Haaker, John P. Scholl Inland Fisheries Jack A. Hanson Wildlife William E. Grenfell, Jr. Editor-in-Chief Robert N. Lea, Ph.D. 131 CONTENTS Page Variation in the Growth Rate of Pacific Herring from San Francisco Bay, California Jerome D. Spratt 132 Migration and Distribution of Northern Pintails Banded in California Warren C. Rienecker 139 Analysis of the Diets of Mountain Sheep from the San Gabriel Mountains, California William M. Perry, Jim W. Dole, and Stephen A. Holl 156 Refutation of Lengths of 1 1 .3, 9.0, and 6.4 m Attributed to the White Shark, Carcharodon carcharias John E. Randall 163 Reproductive Rhythmicity of the Atherinid Fish, Colpichthys regis, from Estero del Soldado, Sonora, Mexico G. A. Russell, D. P. Middaugh, M. J. Hemmer 169 Comparison of Meristic and Morphometric Characters among and within Subspecies of the Sacramento Sucker {Catostomus occidentalis Ayres) David L. Ward and Ronald A. Fritzsche 175 Extent of Human-Bear Interactions in the Backcountry of Yosemite National Park Bruce C. Hastings and Barrie K. Gilbert 188 NOTES First Oregon Record for the Cowcod, Sebastes levis Daniel L. Erickson and Ellen K. Pikitch 192 132 CALIFORNIA FISH AND GAME Calif. Fish and Came 73 ( 3 ): 1 32- 1 38 1 987 VARIATION IN THE GROWTH RATE OF PACIFIC HERRING FROM SAN FRANCISCO BAY, CALIFORNIA1 JEROME D. SPRATT Marine Resources Division California Department of Fish and Game From 1974 to 1982 the average annual increase in body weight of Pacific herring from San Francisco Bay, California was 31%, 22%, 16%, and 9% for ages 3, 4, 5, and 6 yr respectively. In 1983 there was no weight gain and growth in length was 80% be- low normal. Then in 1984, the weight was estimated to have increased 83%, 72%, 53%, and 40% for ages 3, 4, 5, and 6 yr respectively. Growth in length also improved dramatically in 1984. The cessation of growth in 1983 and improvement of growth in 1984 coincided with the beginning and end of the 1983 El Nino. INTRODUCTION Environmental conditions and their effect on fish populations are of primary concern to fishery managers. From late 1982 (Sund 1984) to early 1984 (Fiedler 1984a) the west coast of the United States experienced one of the most severe environmental anomalies ever to occur in the area. Ocean surface temperatures were raised above normal by the intrusion of warm equatorial water; this phe- nomenon is hereafter referred to as El Nino. Age and growth data on Pacific her- ring, Clupea harengus pallasi, from San Francisco Bay California are available from 1974 through 1985 (Spratt 1981, 1982a, 1983a, 1984a, 1985a) and provide an opportunity to study some of the effects that the El Nino had on the growth rate of Pacific herring in California. In 1983, during the height of the El Nino, mean size of herring in California declined drastically (Spratt 1984a). This report doc- uments the variability that is possible in the growth rate of marine fishes. The Pacific herring fishery in San Francisco and Tomales Bays, California has been sampled annually since the 1973-74 season for age and size composition of the purse seine, lampara, and gill net catch. However, only data from the San Francisco Bay purse seine and lampara fisheries were used in this study for the following reasons: (1 ) more data are available from the San Francisco Bay fish- ery, (2) the San Francisco Bay fishery data are continuous from 1974 to 1985, while data from other fisheries are not, and (3) purse seine and lampara fisheries are assumed to be unbiased with regard to size composition of herring in the catch. The San Francisco Bay purse seine and lampara herring fishery is an excellent source of age and growth data. The fishing season is short, with most of the fish caught in a few weeks in January and February. Most fisheries are characterized by irregular pulse fishing that can make annual growth comparisons difficult. The herring fishery is confined to the spawning season, thus occurring at the same point in the herring's life cycle each year. It is believed to be a single stock fishery and minimizes variation caused by stock mixing or fishing multiple stocks. Fi- nally, the purse seine and lampara fishery usually occurs during the latter part of the season when the age composition of the population has a consistent pattern of primarily young herring. This data base consists of 12 yr of data collected 12 'Accepted for publication September 1986. GROWTH RATE OF PACIFIC HERRING 133 months apart each season, and growth estimates are annual, with no need to back calculate or standardize data to a specific month. METHODS A herring sample consisted of 2.3 kg (5 lb) of herring taken with a hand held scoop or by hand from storage boxes. Collecting and processing of samples have remained unchanged since 1973 (Spratt 1981). All samples were processed in the following manner: (i) a 1-kg (2.2-lb) sub-sample was randomly selected from the initial sample; (ii) fish were weighed to the nearest 0.1 g; (iii) body length ( BL) to the nearest mm was measured from the tip of the snout to the end of the silvery part of the body or very near the hypural plate; (iv) sex and ma- turity were determined; and (v) otoliths were removed for age determination. Herring were aged using criteria developed in 1973 (Spratt 1981 ). A total of 2,686 herring was collected from the San Francisco Bay roundhaul fishery from 1974-1985. The Student's t test was used to test for homogeneity between the length and weight of males and females. The condition factor (100,000W/L3) or K was used in this context to estimate the relative well being of the San Francisco Bay herring population. TABLE 1. Mean Length (mm BL) ± Standard Deviation and Number (N) of Herring Sam- pled from the San Francisco Bay Roundhaul Fishery by age and Season. Age Seasonal Unweighted Season 2 3 4 5 6 Mean 1973-74 162±8 174±8 189±8 198±8 206 ±9 186 (152) (81) (65) (53) (20) 1974-75 157±10 179±7 186±6 196±9 205 ±7 185 (52) (24) (16) (13) (12) 1975-76 158±9 174±9 190±6 204 ±10 210±10 187 (75) (51) (15) (17) (17) 1976-77 162 ±9 177 ±9 187±7 198±8 204 ±9 186 (96) (138) (83) (37) (21) 1977-78 162±11 174 ±7 185±9 1 90 ± 11 199±11 182 (34) (30) (32) (11) (2) 1978-79 165 ±7 178±8 186±9 194 ±8 202 ±6 185 (34) (50) (35) (36) (14) 1979-80 160±9 177±8 191 ±7 196 + 8 204 ±8 186 (73) (23) (34) (27) (18) 1980-81 164±6 176±5 188 ±4 197±4 202 ±8 185 (96) (96) (42) (14) (18) 1981-82 163 ±6 175 ±5 184 ±5 192±5 199±4 183 (95) (105) (70) (31) (20) 1982-83 164±5 174 ±5 185±3 194±4 201 ±4 184 (34) (56) (66) (47) (23) 1983-84 152±8 164±6 179±5 188±6 195±4 176 (85) (44) (15) (19) (12) 1984-85 163 ±6 176±6 185±3 193 ±3 198±4 183 (85) (53) (34) (23) (12) Unwt. Mean 161 175 186 195 202 184 RESULTS Length data (Table 1 ) and weight data (Table 2) are summarized by season. Herring up to 9-yr-old were collected but only ages 2 through 6 were used in this analysis because older age classes composed less than 5% of the catch and were 134 CALIFORNIA FISH AND GAME not present in samples every season. Mean length of females age 2 through 6 unweighted for yr class strength was 185 mm, while males averaged 183 mm. Fe- males also weighed more averaging 99 g, while males averaged only 92 g. Neither length nor weight differences were significant (t = —0.192 and —0.625 re- spectively with 4 d.f. ) and data for males and females were combined. The fe- male to male sex ratio of the combined data base was 52:48. Length Analysis From 1973-74 to 1982-83 the combined seasonal average length of herring ages 2 through 6, unweighted for yr class strength, ranged from 182 to 187 mm BL (Table 1 ). In the 1983-84 season the combined unweighted average length decreased to 176 mm BL and then increased to 183 mm BL in the 1984-85 sea- son. In the 1983-84 season, the mean length at age of 2 through 6-yr-old herring was 8 to 12 mm BL below the 10 yr mean length for each age (Table 2). The de- crease in mean length of each age group during the 1983-84 season was not sig- nificant (t = 1.248 with 4 d.f.), nor was the increase in mean length significant in the 1984-85 season. TABLE 2. Comparison of San Francisco Bay Herring 10 Yr Mean Length Data with 1983-84 and 1984-85 Length Data. 10-yr Unwt. Mean mm BL 1983-84 1984-85 Age Mean mm BL mm BL Decrease Mean mm BL mm BL Increase 2 3 4 5 6 162 176 187 196 203 152 ±2.4 164±2.5 179 ±3.6 188 ±3.9 195±3.7 10 12 8 8 8 163 ±1.2 * 176±1.6 185±1.0 193±1.2 198 ±2.4 11 12 6 5 3 Mean 185 176 9.2 183 7.4 * 99% Confidence Interval. A better indicator of the effect of El Nino on growth is obtained if yr classes are followed from season to season. Average annual growth increment was 10 mm for all age classes combined from 1 973-74 to 1 982-83 (Table 3). In 1983 the combined average increase was only 2 mm, well below the 9-yr combined av- erage increase of 10 mm per yr. In 1984, after El Nino ended, the growth rate was above normal with ages 3 through 6 gaining an average of 1 7 mm in length (Table 3). This resurgence in growth was sufficient to allow for a complete recovery in the mean length of herring by the 1984-85 season (Table 1 ). TABLE 3. Annua 1 G rowth Increment of Pacif Age ic herring in mm BL by Age Gro Combined Season 3 4 5 A verage 6 Increment 9 yr ave. 1983-84 1984-85 14 0 24 11 5 21 9 3 14 7 10 1 2 10 17 GROWTH RATE OF PACIFIC HERRING 135 Weight Analysis Average weight at age of herring showed normal variation from the 1973-74 season until the 1982-83 season. During this time the combined seasonal average weight of all age classes, unweighted for yr class strength, ranged between 91 g and 101 g (Table 4). Then in the 1983-84 season, at the height of the El Nino, a severe decline in weight at age occurred that affected adult herring. Environ- mental conditions improved in 1984 with the weakening of El Nino, and in the 1984-85 season herring growth improved dramatically resulting in average weights at age that exceeded long term means for younger age classes (Table 4) . Irrespective of ages, the length-weight relationship of San Francisco Bay herring in 1 983-84 also represents a decrease of up to 7% in weight at length from earlier data (Reilly and Moore 1984). TABLE 4. Mean Weight (g) at Age of Herring in the San Francisco Bay Roundhaul Fishery. Season 1973-74 1974-75 1975-76 1976-77 1977-78 1978-79 1979-80 1980-81 1981-82 1982-83 1983-84 1984-85 Average 61 79 97 113 124 95 The weight loss and subsequent recovery is more evident if year classes are fol- lowed. From 1 975 to 1 983 the average annual weight gain of herring 3 to 6-yr-old was 16 g per year, or an average weight gain of 20% per year (Table 5). The data indicate that during 1983, the same four age classes combined lost an average of 2 g in weight (Table 5). The below average growth increment in 1983 was sig- nificant (t = -4.29 with 3 d.f.). TABLE 5. Annual Growth Increment (g) of Pacific Herring by Age Groups. Age Seasonal Unweighted 2 3 4 5 6 Mean 57 73 95 109. 128 92 55 82 89 105 124 91 57 77 108 122 142 101 58 80 95 116 130 96 66 85 106 114 111 96 70 87 103 119 128 101 68 79 102 119 132 100 63 83 98 124 118 97 61 82 98 113 124 96 62 74 93 107 120 91 46 58 75 91 103 75 66 84 100 115 127 98 Age Combined Season 3 4 5 6 Average 9 yr ave. Grams 19 18 16 10 16 Percent 31 22 16 9 20 1983-84 Grams -4 1 -2 -4 -2 Percent -6 1 -2 -4 -3 1984-85 Grams 38 42 40 36 39 Percent 83 72 53 40 62 At the end of the 1983-84 season, herring were underweight by at least 20%. When conditions improved in 1984, herring rapidly regained weight that was lost and by the 1984-85 fishery, herring were robust with average weights at age above long term means. Three-yr old herring gained 38 g on the average during 136 CALIFORNIA FISH AND GAME 1984 for a total increase in weight of 83%; 4, 5, and 6-yr-olds also had large in- creases in average weight. The combined average increase in weight of 39 g was more than double the long term average increase and represents a 62% average increase in weight during one season (Table 5). Condition Factor From the 1 973-74 to 1 984-85 season, K averaged 1 .48 for age groups 2 through 6 unweighted for yr class strength (Table 6). The San Francisco Bay herring biomass estimates increased from 1978-79 to 1981-82, peaking at 100,000 tons (Spratt 1981, 1982b), then began a decline to 40,000 tons in the 1983-84 season (Spratt 1983b, 1984b). The increase in biomass was partly due to improved methodology (i.e. including subtidal spawn- ing), but an actual increase in biomass probably occurred because the 1978, 1979, and 1980 yr classes were all above average strength (Spratt 1981, 1982a). During the four seasons that the herring biomass was increasing the annual K val- ues were above average, indicating that herring were in relatively good condition at that time (Table 6). When the biomass declined the K values were below av- erage (Table 6), indicating the relatively poor condition of herring during those years. Herring biomass increased to 46,000 tons in the 1984-85 season (Spratt 1985b), coinciding with the weakening of El Nino and above average K values (Table 6). TABLE 6. Condition Factor (100,000W/L3) for San Francisco Bay Herring, calculated from Ta- bles 1 and 4. Season ]973-74 1974-75 1975-76 1976-77 1977-78 1978-79 1979-80 1980-81 1981-82 1982-83 1983-84 1984-65 Mean 1.45 1.46 1.49 1.51 1.49 1.48 DISCUSSION Length Growth in length slowed measurably in 1983 and then made a complete re- covery in 1984. Reilly and Moore (1984, 1985) in an independent study, found that the average length at age of adult herring in San Francisco Bay was also be- low normal in 1983 and recovered in 1984. My data indicate a greater decline oc- curred, but the two studies do agree that growth was poor in 1983. Differences in growth estimates could be related to sample size, sampling gear, time of sam- pling, or aging differences. Similar poor growth was noted for northern anchovy, Engraulis mordax, in southern California during 1 983. Feidler ( 1 984a ) found that adult anchovies were Age Seasonal unweighted 2 3 4 5 6 mean 1.34 1.38 1.40 1.40 1.46 1.40 1.42 1.42 1.38 1.39 1.43 1.41 1.44 1.46 1.57 1.43 1.53 1.49 1.36 1.44 1.45 1.49 1.53 1.45 1.55 1.61 1.67 1.66 1.40 1.58 1.55 1.54 1.60 1.61 1.55 1.57 1.66 1.42 1.46 1.58 1.55 1.53 1.42 1.52 1.47 1.62 1.43 1.49 1.40 1.53 1.57 1.59 1.57 1.53 1.40 1.40 1.46 1.46 1.47 1.44 1.30 1.31 1.30 1.36 1.38 1.33 1.52 1.54 1.57 1.59 1.63 1.57 GROWTH RATE OF PACIFIC HERRING 137 abnormally small, with a mode of 100-105 mm compared to an expected mode of 115-120 mm. Weight The average weight at age of herring in San Francisco Bay decreased to an all time low of 75 g in 1983 (Table 4). While the data indicate that some individual herring lost weight in 1983, I believe that the overall population merely main- tained their weight from 1982 to 1983. Poor growth was related to low primary productivity during the El Nino (Fiedler 1984b). Reilly and Moore (1984) found that herring fecundity did not decline significantly during the El Nino, suggesting that available energy was directed to gonadal development rather than growth. Other recent work on central California fishes indicate that growth of blue rockfish, Sebastes mystinus, was also below normal in 1983. The weight-length relationship of adult females in 1983 was 20% below 1981 levels, strongly sug- gesting that blue rockfish may have lost weight in 1983 ( D. VenTresca, Calif. Fish and Game, pers. comm.). Population Growth Rate Factors that cause changes in the growth rate of herring populations are either environmental, effecting primary productivity, or are related to population size and are density dependent. Both factors influenced herring growth in 1983 and 1984. The degree of change in the annual growth rate of a population is depen- dent on the age composition, because younger fish gain weight faster than older fish. Since the San Francisco Bay herring population is weighted toward younger age classes, it should respond significantly to factors that effect growth. Accord- ingly, the annual average population growth rate (weight gain) will be closer to the long term average increase for 3-yr-olds (31%) than the combined average increase of 20% for all age classes (Table 5). Similarly, the herring population growth rate in 1984 was probably closer to the increase exhibited by 3-yr-olds (83%) in 1984 than the average of 62% for all age classes combined (Table 5). When the annual growth rate of a fish population fluctuates between zero and approximately 70%, growth or lack of growth can have a significant effect on biomass. CONCLUSION This study shows that there is wide variation in the San Francisco Bay herring population growth rate and indicates that the annual growth rate (weight) ranges between zero and 0.7 with a mean of about 0.3. The weight at age of herring and the overall mean weight of the catch was depressed in 1983 and then recovered quickly in 1984. Therefore the weight gain in 1984 represents the maximum re- corded growth rate for San Francisco Bay herring. The 1983 and 1984 growth fluctuations coincided with the 1983-84 El Nino event. LITERATURE CITED Feidler, Paul C. 1984a. Some Effects of El Nino 1983 on the Northern Anchovy. CALCOFI Rep., 25:53-58. 1984b. Satellite Observations of El Nino along the U.S. Pacific Coast. Science, 224:1251-1254. Reilly, P. N. and T. O. Moore, 1984. Pacific Herring, Clupea harengus pallasi. Studies in San Francisco Bay, Monterey Bay, and the Gulf of the Farallones, May 1983 to March 1984. Calif. Dept. Fish and Came, Mar. Re- sources Admin. Rept. 84-3:1-67. 1985. Pacific Herring, Clupea harengus pallasi, Studies in San Francisco Bay and the Gulf of the Faral- lones, June 1984 to March 1985. Calif. Dept. Fish and Game, Mar. Resources Admin. Rept. 85-4:1-73. 1 38 CALIFORNIA FISH AND GAME Spratt, J. D. 1981. Status of the Pacific herring Clupea harengus pallasi, in California to 1980. Calif. Fish and Came, Fish Bull. 171:1-104 1982. Biological Characteristics of the Catch from the 1980-81 and 1981-82 Pacific herring, Clupea harengus pallasi, Roe Fishery in California. Calif. Fish and Came, Mar. Resources Admin. Rept. 82-10:1-16. 1982b. Biomass Estimates of Pacific Herring. Clupea harengus pallasi, in California from the 1981-82 Spawning Ground Surveys. Calif. Dept. Fish and Came, Mar. Resources Admin. Rept. 82-6:1-19. 1983. Biological Characteristics of the Catch from the 1982-83 Pacific herring, Clupea harengus pallasi, Roe Fishery in California. Calif. Fish and Came, Mar. Resources Admin. Rept. 83-4:1-11. 1983b. Biomass Estimates of Pacific Herring, Clupea harengus pallasi, in California from the 1982-83 Spawning Ground Surveys. Calif. Dept. Fish and Game, Mar. Resources Admin. Rept. 83-3:1-23. , 1984. Biological Characteristics of the Catch from the 1983-84 Pacific herring, Clupea harengus pallasi, Roe Fishery in California. Calif. Fish and Game, Mar. Resources Admin. Rept. 84-4:1-18. 1984b. Biomass Estimates of Pacific Herring. Clupea harengus pallasi, in California from the 1983-84 Spawning Ground Surveys. Calif. Dept. Fish and Game, Mar. Resources Admin. Rept. 84-2:1-29. 1985a. Biological Characteristics of the Catch from the 1984-85 Pacific herring. Clupea harengus pallasi, Roe Fishery in California. Calif. Fish and Game. Mar. Resources Admin. Rept. 85-3:1-24. 1985b. Biomass Estimates of Pacific Herring, Clupea harengus pallasi, in California from the 1984-85 Spawning Ground Surveys. Calif. Dept. Fish and Game, Mar. Resources Admin. Rept. 85-2:1-31. Sund, Paul N. 1984. Atlas of Airborne Sea Surface Temperature Observations in Nearshore California Waters, 1980-1983. NOAA Technical Memorandum, SFWC, 43:1-102. MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 1 39 Calif. Fish and Game 73 ( 3 ): 1 39-1 55 1 987 MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS BANDED IN CALIFORNIA1 WARREN C. RIENECKER California Department of Fish and Game 1416 Ninth Street Sacramento, California 95814 Migration routes and distribution of northern pintails, Anas acuta, were deter- mined by analysis of band recoveries from 234,061 pintails banded at 12 stations in California from 1948 to 1979. Males were the first to arrive in California during fall mi- gration resulting in a differential distribution of harvest. Females were harvested closer to the breeding grounds than males, whereas males were harvested farther south. Males tended to range wider and were more apt to be recovered in Mexico and Central America or one of the other three flyways than were females. Females showed a greater homing instinct to the area of banding than did males. Propor- tionately fewer bands were recovered in the Sacramento Valley from preseason ver- sus postseason banding at Gray Lodge, suggesting that some preseason banded pintails were only passing through the Sacramento Valley on their way farther south. The large number of local recoveries from the San Francisco Bay-Delta, especially Suisun Marsh, suggest this area was a highly desirable pintail wintering area in the 1950's. Los Banos banded pintails were more closely aligned with Mexico and the other flyways than were birds banded in other areas of northern California. Imperial Valley pintails were a separate population from those of northern California. Pintails banded preseason in the Imperial Valley wintered in four general areas: Imperial Val- ley, northern California, Mexico and in the Central and Mississippi Flyways. Some pintails that migrated to Mexico returned to the breeding ground via the Central and Mississippi flyways on a counterclockwise migration, whereas others migrated up through California on a clockwise migration. Thus, several subpopulations were banded at each station partially accounting for the spread of recoveries. INTRODUCTION An average of two million northern pintails winter in California (Bellrose 1976). About 71% of the U.S. harvest occurs in the Pacific Flyway, and over half in California (Geis and Cooch 1972) . Texas and Louisiana are the only significant harvest areas outside of the Pacific Flyway. Most pintails wintering in the Pacific Flyway are produced in prairie Canada and Alaska (Bellrose 1976). Pintails banded in Alaska show 70% being recovered in the Pacific Flyway, and 41% of the Pacific Flyway recoveries came from California (Bird Banding Laboratory, unpubl. data). Pintails were harvested in larger numbers than other species of ducks in California. The purposes of this study were to describe migration patterns and distribution of pintails banded pre- and postseason in California to provide a basis for man- agement and to determine the relative importance of California as a wintering area and harvest area for California banded pintails. METHODS Pintails were banded by the Waterfowl Studies Project of the California De- partment of Fish and Game (CDFG ), in cooperation with the U.S. Fish and Wild- 1 A contribution of Federal Aid in Wildlife Restoration Project W-30-R, "Waterfowl Studies," accepted for pub- lication August 1986. 140 CALIFORNIA FISH AND CAME life Service (USFWS) during 1948-79. Most pintails were caught in swim in baited wire traps, but cannon nets were also used postseason at Gray Lodge Wildlife Area and in Imperial Valley. Of total ducks banded, 175,985 were preseason, 69,189 postseason, 4,371 within season and 6,008 between split sea- sons. However, only those banded preseason (174,896) and postseason (59,165) that were recovered through the 1979 band recovery year were used. In the 1950's and early 1960's, pintails were banded on several waterfowl con- centration areas in California. However, from 1964 to 1979, only the Klamath Ba- sin NWRs and the Gray Lodge Wildlife Area were used as pintail banding sta- tions. Banding on these two areas has been continuous since 1948. Recoveries from each banding station, when warranted, were analyzed sep- arately. However, data from stations with similar recovery distributions were pooled. Data from each age and/or sex class were pooled where there were no obvious differences noted. All percentages used in this report were expressed as proportions of total recoveries resulting from a particular banded sample. Direct and indirect recoveries were used. Direct recoveries are banded birds killed or found dead during the first hunting season after banding. Indirect re- coveries are band recoveries in subsequent band recovery years following the year of banding. All recoveries were from normal wild birds in banded samples of 100 or more. California was divided into eight band recovery areas (Figure 1 ). Each area represents a major subdivision separated by geographic features, waterfowl dis- tribution and harvest. The most important were northeastern California; Sacra- mento, San Joaquin and Imperial Valleys; and San Francisco Bay-Delta as deter- mined from waterfowl populations using the areas. RESULTS AND DISCUSSION Migration, Bandings and Recoveries Klamath Basin NWRs (Northeastern Recovery Area): Banded 52,130 preseason, 1948-1979. Distribution of direct recoveries of immature males and females (Table 1) banded on the Klamath Basin NWRs were similar. Recoveries of adults differed for northeastern California, Sacramento Valley, and San Franciso Bay-Delta re- covery areas and were treated separately. There was a predominance of males during the August and September preseason banding in the Klamath Basin; 43,306 males and 8,824 females were banded. By the time hunting began in mid-October, many adult males had moved south, so fewer were shot in the Klamath Basin compared to later mi- grating adult females. Consequently, a higher proportion of band recoveries were obtained for females (20.9%) from the Klamath Basin than for males (12.6%). This was shown in both direct and indirect band recoveries (Table 1 ). Delayed female migration was also shown by the proportion of indirect recoveries from Canada, where 9.8% of the females were recovered compared to 3.6% for males, and also by the number of recoveries obtained in Washington and Oregon (females 9.4% vs. males 5.2%). Bellrose et al. (1961) stated that excess hen pintails in the hunter bag in Manitoba was caused by the dispersal of many adult drakes away from the area prior to the hunting season. As a result of differential MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 141 KLAMATH BASIN HUMBOLDT BAY MT. MEADOWS GRAY LODGE PT. REYES SUISUN SO. S. F. BAY LOS BANOS DUNES LAKE PT. MUGU IMPERIAL VALLEY LEGEND BANDING STATIONS •- RECOVERY AREA BOUNOARY FIGURE 1. Pintail Banding Stations and Recovery Areas. migration, indirect recoveries of males exceeded those of femalesin the Sacra- mento Valley (27.9% vs. 24.5%) and San Francisco Bay-Delta (28.6% vs. 14.7%). The data suggest that most females had fixed migration routes from which they rarely deviate, whereas males were more apt to stray from one route to another. Over 80% of male and 70% of female indirect recoveries of pintails banded on Klamath Basin NWRs occurred in California (Table 1 ). Thus, pintails migrated to California prior to the opening of the hunting season or with a minimum of stopovers along the way. 142 CALIFORNIA FISH AND GAME TABLE 1. Distribution of Band Recoveries (Percent of Recoveries) From 29,922 Adult Male, 13, 384 Immature Male, 3,910 Adult Female and 4,914 Immature Female Pintails Banded Preseason at Klamath Basin National Wildlife Refuges, 1948-79 Direct /mm. male Adult male /mm. female Adult female Indirect* Recovery Areas* Males Females California Northeast 34.8 21.5 33.7 44.2 12.6 20.9 North Coast 1.1 0.9 0.7 1.8 1.0 0.9 Sacramento Valley 15.3 25.1 17.4 17.7 27.9 24.5 San Francisco Bay 22.3 27.4 20.0 17.7 28.6 14.7 South Coast 3.0 1.8 2.6 3.5 1.1 1.5 San Joaquin Valley 11.1 13.0 13.0 11.5 11.5 10.1 Imperial 0.9 0.1 0.4 - 0.4 1.2 Washington Coast 1.3 0.8 2.6 - 1.1 2.1 Oregon Coast 3.8 2.5 3.0 - 2.4 3.4 South 1.0 0.6 1.1 1.8 1.2 1.8 Nevada - 0.2 1.1 - 0.5 0.3 Utah 0.1 0.5 - - 0.8 0.9 Alaska 0.2 - - - 0.9 0.6 British Columbia 0.5 0.3 1.9 - 0.5 1.8 Alberta - 0.2 - 0.9 1.9 4.9 Saskatchewan - 0.2 - - 0.9 2.8 Mexico 2.3 1.4 1.1 - 1.5 1.8 Central Flyway 1.2 2.0 0.4 - 2.5 2.7 Mississippi Flyway 0.3 1.3 - - 0.9 0.3 Russia - 0.1 - - 0.8 0.3 Oceana 0.1 - 0.4 0.9 - - All Other* 0.8 0.5 1.2 1.2 2.1 Total Recoveries 1006 1138 270 113 3538 326 * Table includes data for recovery percentages > 0.7 only. + Birds in their second year or older. * Recovery areas with < 0.7 To determine if the distribution of band recoveries changed during the study, Klamath Basin data for the 1950's, 1960's, and 1970's were compared (Table 2). Combining sexes, there was a downward trend in recoveries from the San Francisco Bay-Delta Recovery Area (29.2%, 28.6%, 23.9%) and Canada (6.2%, 3.1%, 2.3%), and an upward trend for recoveries from the Sacramento Valley (24.7%, 27.3%, 31.8%) and northeastern California (12.7%, 10.3%, 17.9%). Except for females recovered in the San Francisco Bay-Delta, the same trends oc- curred when sexes were treated separately. Later migrations may have resulted in more banded birds being available to hunting in the Klamath Basin. In the 1950's and early 1960's, pintails were banded on Klamath Basin NWRs in August. However, by the late 1960's, early movement of pintails into the Klamath Basin declined and was delayed by about one month. Consequently, banding was not done until September. Thus, more banded birds were on the area at a later date. Increasing numbers of pintails over-wintering in Klamath Basin (H. McCollum, pers. commun.) could also contribute to the in- creased recoveries. The cause of fewer recoveries from the San Francisco Bay- Delta Recovery-Area may reflect the change in northeastern California. In 1957, on the Klamath Basin NWRs, two preseason samples of pintails were banded. The early sample ( N = 1 963 ) was banded in August, and the late sample ( N = 1 ,697 ) was banded just prior to the October opening of hunting season. The only noticeable difference in distribution of recoveries between the two samples 12.7 10.3 17.9 0.8 0.7 1.3 24.7 27.3 31.8 29.2 28.6 23.9 1.4 1.0 0.9 9.6 14.1 11.0 1.2 0.8 0.5 1.0 0.8 0.7 3.1 1.7 1.4 1.9 0.8 0.5 1.8 1.6 1.3 4.1 1.9 1.2 8.7 10.4 7.6 MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 143 TABLE 2. A Comparison of Indirect Band Recoveries (percent of recoveries) Among 3 De- cades of 48,121 Pintails Banded Preseason at the Klamath Basin National Wildlife Refuges, 1950-79 Recovery Area * 1950-59 1960-69 1970-79 California Northeast North Coast Sacramento Valley San Francisco Bay South Coast San Joaquin Valley Utah Alaska Alberta Saskatchewan Mexico Central Flyway All Other+ Total Recoveries 1455 1343 985 * Table includes data from recovery percentages > 0.7 only. Complete list available from au- thor. + Recovery areas with < 0.7. was the number of local recoveries (August 1 1.8%; October 30.7%). Direct re- coveries outside the State were about the same for both periods which suggests that both samples were from the same population. Honey Lake Wildlife Area and Mountain Meadows (Northeastern Recovery Area): Banded 20,638 preseason, 1950-1958. Recovery locations of pintails banded at Mountain Meadows and Honey Lake Valley were no different, so data were pooled (Table 3) . There were differences in recovery areas between Honey Lake and Klamath Basin banding stations even though the stations are only 240 km apart. Local indirect recoveries in North- eastern California from Honey Lake — Mountain Meadow birds was 9.0% (Table 3) compared to 12.6% (Table 1 ) for Klamath Basin birds. This same pattern held for direct recoveries of immatures. Thus, some of these pintails used different mi- gration routes or there was greater hunting pressure in the Klamath Basin. Fewer Honey Lake than Klamath Basin banded pintails were recovered in the Sacramento Valley (18.8% vs. 25.8%) and San Francisco Bay-Delta (27.7% vs. 31.0%). More were recovered in the San Joaquin Valley (16.2% vs. 9.6%), southern California (3.2% vs. 1.9%), Mexico (3.3% vs. 2.0%) and other flyways (7.5% vs. 4.8%). Gray Lodge Wildlife Area (Sacramento Valley Recovery Area): Banded 68,584 preseason and postseason, 1949-1979. The proportional distributions of direct recoveries were the same for immature males and females (Table 4), as in the Klamath Basin. Immatures tended to dis- perse farther than did adults (Table 4). Direct recoveries in the Sacramento Val- ley were high compared to indirect recoveries (Table 4) . This difference was ex- pected since pintails were banded shortly before the hunting season. Males showed more proportionate recoveries from the San Francisco Bay- Delta (31 .7%) than did females (20.4%) and fewer from the Sacramento Valley (36.7% vs 40.7%) respectively (Table 4). Other differences in band recoveries such as 2.8% for males in northeastern California compared to 6.0% for females 144 CALIFORNIA FISH AND GAME TABLE 3. Distribution of Band Recoveries (percent of recoveries) from 20,638 Pintails Banded Preseason at Honey Lake Wildlife Area and at Mountain Meadows, 1950-58 Direct Recovery Area* Immature Adult Indirect* California Northeast 23.9 15.3 9.0 Sacramento Valley 14.2 14.5 18.8 San Francisco Bay-Delta 26.8 34.8 27.7 South Coast 2.6 3.2 2.3 San Joaquin Valley 19.1 20.8 16.2 Imperial Valley 0.8 1.1 0.9 Washington 0.8 0.3 1.3 Oregon 1.8 0.8 2.5 Nevada 0.4 1.3 1.2 Utah 0.1 0.5 1.2 Alaska 0.1 0.3 0.7 Alberta 0.1 0.5 3.0 Saskatchewan - - 1.6 Mexico 4.4 1.1 3.3 Central Flyway 1.8 4.2 5.5 Mississippi Flyway 1.6 1.1 2.0 Russia 0.7 All Other * 1.4 0.3 2.1 Total Recoveries 754 379 1420 * Table includes data from recovery percentages >0.7 only. Complete list available from author. Birds in their second year or older. ^ Recovery areas with <0.7. and 4.7% for males on the breeding grounds of Alaska and Canada compared to 9.2% for females show later southward migration of females. A greater proportion of indirect recoveries of pintails banded preseason at Gray Lodge come from south and east of California than recoveries from postseason banded birds ( Tables 4 and 5 ) . Thus, some of the preseason banded birds were only passing through the area on their way farther south; whereas, birds banded postseason were on their wintering grounds or migrating north. Preseason direct recoveries also indicate this continuation of the migration south and east after banding (Table 4). Preseason versus postseason bandings at Gray Lodge also showed a marked difference in the distribution of band recoveries within California. About 38% of pintails banded preseason were recovered in the Sacramento Valley compared to 46.3% from postseason bandings. There was also a greater dispersal of pintails from preseason bandings compared to postseason bandings. The difference in number of indirect recoveries within northeastern California from preseason (male 2.8% vs. female 6.0%) and postseason bandings (male 6.3 vs. female 7.0%) at Gray Lodge resulted from differential migration (Tables 4 and 5). Males causesd most of this variation (preseason 2.8% vs. postseason 6.3%), whereas females were similar (preseason 6.0% vs. postseason 7.0%). Thus, preseason banded pintails, especially males, migrated earlier than those birds banded postseason. They passed through northeastern California prior to or shortly after opening of the hunting season, thereby resulting in fewer bands be- ing recovered for that area. MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 145 TABLE 4. Distribution of Band Recoveries (percent of recoveries) from 15,446 Adult Male, 11,824 Immature Male, 7,535 Adult Female and 9,500 Immature Female Pintails Banded Preseason at Gray Lodge Wildlife Area, 1949-79 Direct Indirect + /mm. Adult Imm. Adult Recovery Areas * male male female female Males Females California Northeast 1.2 - 0.6 0.5 2.8 6.0 Sacramento Valley 54.5 49.8 54.4 59.7 36.7 40.7 San Francisco Bay-Delta 28.5 32.8 25.9 28.3 31.7 20.4 South Coast 0.8 1.0 1.4 1.0 1.2 1.2 San Joaquin Valley 11.1 13.4 14.3 10.5 12.2 10.8 Imperial Valley 0.5 0.3 0.4 - 0.8 0.7 Washington - - 0.1 - 0.8 0.9 Oregon 0.4 0.3 0.8 - 2.9 3.4 Utah - - - - 0.8 1.2 Alaska - 0.2 0.2 - 1.2 1.6 British Columbia - 0.3 - - 0.7 0.6 Alberta - - - - 1.7 5.3 Saskatchewan - - - - 0.8 1.7 Mexico 1.4 0.2 1.0 - 1.2 0.6 Central Flyway 0.7 1.5 0.4 - 1.7 0.5 Russia - - - - 0.7 0.9 All Other^ 0.6 - 0.5 - 2.3 3.1 Total Recoveries 943 618 491 191 2356 813 * Table includes data for recovery percentages > 0.7 only. Complete list available from author. + Birds in their second year or older. ^ Recovery areas with < 0.7. TABLE 5. Distribution of Indirect Band Recoveries (percent of recoveries) from 15,672 Male and 8,607 Female Pintails Banded Postseason at Gray Lodge Wildlife Area, 1954-79 Recovery Areas * California Northeast Sacramento Valley San Francisco Bay-Delta San Joaquin Valley Imperial Valley Washington Oregon Nevada Utah Alaska British Columbia Alberta Saskatchewan Central Flyway Russia All Other + Total Recoveries 1438 416 1854 * Table includes data from recovery percentages > 0.7 only. Complete list available from author. + Recovery areas with > 0.7. Vlale Female Combined 6.3 7.0 6.4 44.2 53.6 46.3 22.7 10.1 19.9 12.7 10.6 12.2 0.7 - 0.5 0.8 2.2 1.1 3.1 4.3 3.4 1.1 2.2 1.3 1.0 1.4 1.2 0.8 1.0 0.8 0.8 1.9 1.0 2.2 1.4 2.0 0.5 1.4 0.9 0.4 0.7 0.5 0.6 1.0 0.7 2.1 1.2 1.8 146 CALIFORNIA FISH AND CAME Band recoveries for Gray Lodge bandings were also compared for the 1950's, 1%0's and 1970's (Table 6). There was an increase in percentage of recoveries from the Sacramento Valley during the 1950's through the 1960's (33.2%— 42.4%) and a leveling off in the 1970's (40.2%). Increases were also shown for the San Francisco Bay-Delta (26.2%, 30.4%, 32.8%) and San Joaquin Valley (10.4%, 11.7%, 13.6%), whereas, there was a downward trend in Canada (7.5%, 2.3%, 1.9%). TABLE 6. A Comparison of Indirect Band Recoveries (percent of recoveries) Among 3 De- cades of .39,504 Pintails Banded Preseason at Gray Lodge Wildlife Area, 1950-79 Recovery Areas * 1950-59 1960-69 1970-79 California Northeast Sacramento Valley San Francisco Bay-Delta South Coast San Joaquin Valley Oregon Utah Alaska Alberta Saskatchewan Mexico Central Flyway Mississippi Flyway Russia All Other+ Total Recoveries 1507 694 938 * Table includes data from recovery percentages >0.7 only. Complete list available from author. + Recovery areas with <0.7 Point Reyes, Suisun, and South San Francisco Bay (San Francisco Bav-Delta Recovery Area): Banded Suisun, 8,715; South San Francisco Bay, 4,089; and Point Reyes, 1,589 preseason, 1951-1958. Recoveries from pintails banded on these stations had similar distribution so data were pooled (Table 7). These three banding stations had a higher propor- tion of local direct (adults 79.6%, immatures 74.4%) and indirect (58.6%) re- coveries than found for Gray Lodge (52.2%, 54.5%, 37.7% respectively; Table 4). Once pintails arrived in the San Francisco Bay-Delta area, they remained there. Those banded on other areas of the wintering grounds had a greater tendency to disperse. Coastal Banding Stations (North Coast and South Coast Recovery Areas): Banded Humboldt Bay, 121; Dunes Lake, 535; and Point Mugu, 813 preseason, 1953-1958. Recoveries from the coastal stations did not reveal any different migration pat- terns than were found for the inland stations. Of the 85 indirect recoveries, two might be regarded as coming from the direction of the Canada breeding grounds (one was from Canada) and four from the direction of Alaska (one from Sibe- ria). Direct recoveries indicated that some continued on to the Central Flyway and Mexico. 4.8 1.4 3.3 33.2 42.4 40.3 26.2 30.4 32.8 1.7 0.9 0.3 10.4 11.7 13.6 3.8 3.3 1.1 1.1 0.4 0.4 1.3 1.2 1.5 4.2 1.7 0.9 1.9 - 0.5 1.8 0.2 0.3 2.9 0.8 0.6 0.8 0.1 0.3 0.3 1.3 1.1 5.7 3.9 2.6 MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 147 TABLE 7. Distribution of Band Recoveries (percent of recoveries) From 14,393 Pintails Banded Preseason at Point Reyes, Suisun and South San Francisco Bay, 1951-58 Direct Indirect* Recovery Areas * Immature Adult California Northeast 0.4 - 2.6 Sacramento Valley 10.2 9.9 15.7 San Francisco Bay-Delta 74.3 79.6 58.6 South Coast 1.4 0.5 1.0 San Joaquin Valley 10.2 7.0 9.9 Washington 0.4 - 0.8 Oregon - 0.2 1.2 Utah - 0.2 0.8 Alberta - 0.2 2.4 Saskatchewan - 0.4 0.9 Mexico 1.1 0.4 0.7 Central Flyway 0.8 0.7 1.4 All Others * 1.2 1.0 3.9 Total Recoveries 1007 554 1096 * Table includes data from recovery percentages > 0.7 only. Complete list available from author. + Birds in their second year or older. ** Recovery areas with < 0.7 Los Banos (San Joaquin Valley Recovery Area) : Banded 28,623 preseason and postseason, 1948-1962. Much of the information already learned from the other banding stations also applies to Los Banos ( Tables 8 and 9 ) , e.g., the delayed migration of females from the breeding grounds, the greater number of females than males recovered in the area of banding, and the greater percentage of recoveries from postseason bandings in the area of banding compared to preseason bandings. San Francisco Bay-Delta Recovery Area was an important harvest area for males banded at Los Banos. In fact, more indirect recoveries of Los Banos males banded preseason were recovered in the San Francisco Bay-Delta (30.7%) than in the San Joaquin Valley (25.8%) where they were banded (Table 8). However, recoveries from postseason banding show more males recovered in the San Joaquin Valley (35.7%) than in the San Francisco Bay-Delta (25.7%, Table 9). Direct recoveries suggested immature males wandered farther from the band- ing area than did adult males and females. For example, 16.4% of direct recov- eries for immature males were recovered outside of California compared to 10.1% for adult males, 7.0% for immature females and 1.1% for adult females (Table 9) . This difference was greater than among Gray Lodge pintails (3.1% im- mature males, 3.0% immature females and 0.0% adult females) recovered out- side of the State (Table 4). Mexico and the Central and Mississippi flyways ac- counted for the majority of these out of state recoveries (14.2% immature males, 4.0% immature females). Thus, Los Banos banded pintails were more closely aligned with Mexico and the other flyways than were birds banded in other areas of northern California. Imperial Valley (Imperial Valley Recovery Area): Banded 48,228 preseason and postseason, 1951-1973. Reports on lesser snow geese, Anserc. caerulescens ( Rienecker 1965), Amer- ican wigeon, Anser americana (Rienecker 1976), and green-winged teal, Anser 148 CALIFORNIA FISH AND GAME crecca carolinensis (Moisan et al. 1967), showed that species which winter in the Imperial Valley were different populations than those wintering in northern California. The distribution of pintails banded in the Imperial Valley followed this pattern. Most indirect recoveries from postseason banded pintails occurred in the Imperial Valley (41.1%, Table 10). Utah was second in number of recoveries (14.2%); thus, pintails stopped in Utah before continuing into the Imperial Val- ley. Also, 21 .5% of the postseason banded pintails were recovered in the Central Valley of California. The peak month of band recoveries, both direct and indirect, was October for the San Joaquin Valley, November for the San Francisco Bay- Delta, and November-December for the Sacramento Valley (Table 11 ). As shown by the northern California bandings, winter banded females showed greater homing to the area of banding than did males. About 46% of the females were recovered in Imperial Valley from postseason banding compared to 38.9% for males. Preseason bandings show a much higher proportion of local indirect recoveries among females (34.8%) than males (16.3%, Table 12), suggesting that males dispersed widely. More Imperial Valley preseason banded male pintails were harvested in Mex- ico and Central America (10.3%) than for males banded preseason at Gray Lodge ( 1 .3%, Table 4) . This is reasonable since Imperial Valley borders on Mex- ico. There were marked differences between Imperial Valley preseason and postseason bandings. Fewer bands were recovered from the recovery area where banded when banding occurred preseason compared to postseason. For example, pintails that were banded in Imperial Valley had recoveries of males TABLE 8. Distribution of Band Recoveries (percent of recoveries) From 9,961 Adult Male, 7,891 Immature Male, 2,090 Adult Female and 4,913 Immature Female Pintails Banded Preseason at Los Banos, 1948-62 Direct Indirect /mm. Adult Imm. Adult Recovery Areas * male male female female Males Females California Northeast 0.8 0.3 0.3 - 3.4 1.9 Sacramento Valley 13.2 11.8 11.4 8.5 15.2 13.6 San Francisco Bay-Delta 27.1 26.3 30.2 23.4 30.7 23.0 South Coast 4.2 2.5 2.2 3.2 1.5 1.1 San Joaquin Valley 34.9 48.0 48.5 62.8 25.8 36.3 Imperial Valley 2.3 0.5 - 1.1 1.3 1.1 Oregon 0.7 0.2 1.1 - 0.5 1.4 Nevada 0.3 0.2 0.3 - 1.1 1.9 Utah - - - - 1.1 3.6 Alaska 0.2 0.2 - - 0.9 0.3 Alberta - - - - 1.9 5.5 Saskatchewan - - - - 1.5 0.6 Mexico 7.2 2.6 2.4 - 3.6 3.5 Central Flyway 6.7 4.7 1.1 1.1 7.3 1.7 Mississippi Flyway 0.3 1.0 0.5 - 1.9 1.4 All Other * 2.1 1.7 2.0 - 2.3 3.1 Total Recoveries 598 598 367 94 1498 361 * Table includes data for recovery percentages > 0.7 only. Complete list is available from author. + Birds in their second year or older. ^ Recovery areas with < 0.7. MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 149 16.5% vs. females 34.8% for preseason and males 38.9% vs. females 46.2% for postseason bandings within the same area. Thus, most pintails banded preseason in Imperial Valley did not winter there but were only passing through on their fall migration to the wintering grounds. TABLE 9. Distribution of Indirect Band I Recoveries (pi jrcent of recoveries] i Froi n 2,27; and 1,491 Female Pintails Banded Preseason at Los Banos, 1953-59 Recovery Areas * Male Female Combined California Northeast 3.4 2.6 3.2 Sacramento Valley 18.6 13.8 17.5 San Francisco Bay-Delta 25.7 11.2 22.3 South Coast 2.9 2.6 2.8 San Joaquin Valley 35.7 50.0 39.0 Washington 0.9 - 0.6 Oregon 1.0 0.9 1.0 Nevada 0.8 0.9 0.8 Utah 2.1 2.6 2.2 Alaska 1.3 - 1.0 Alberta 1.6 2.6 1.8 Saskatchewan 1.0 2.6 1.4 Mexico 0.5 2.6 1.0 Central Flyway 1.3 2.6 1.6 All Others + 2.9 4.6 3.6 Total Recoveries 381 116 497 Table includes data from recovery percentages > 0.7 only. Complete list available from author. " Recovery areas with < 0.7. TABLE 10. Distribution of Indirect Band Recoveries (percent of recoveries) From 17,538 Male and 13,595 Female Pintails Banded Postseason at Imperial Valley, 1952-73 Recovery Areas * California Northeast Sacramento Valley San Francisco Bay-Delta South Coast San Joaquin Valley Imperial Valley Nevada Utah Arizona Alberta Saskatchewan Canada (Other) Mexico Central Flyway Russia All Others+ Total Recoveries 1914 833 2747 * Table includes data from recovery percentages > 0.7 only. Complete list available from author. + Recovery areas with < 0.7. vlale Female Combined 1.0 1.8 1.3 4.3 2.9 3.9 5.6 4.3 5.2 2.0 2.4 2.1 14.9 6.5 12.1 38.9 46.2 41.1 2.1 2.2 2.1 14.7 13.1 14.2 1.4 1.3 1.5 1.7 2.8 2.0 1.0 1.8 1.3 0.1 0.7 0.3 7.2 6.7 6.9 2.1 2.4 2.2 0.5 0.7 0.5 2.8 3.9 2.7 1 50 CALIFORNIA FISH AND CAME "8 i— u"l CM «— Iv. u"> l/l ff\ CO ^B n ffi o^ "— ■— »— co i— ^ o CTi ro »— »— no -Cj P o^ tt © iv. r-s ^ N O CO oS (N i-I cf diNOtmo CM i— CO vO vO 00 O «— CM CM 00 _ (N »_ O T ^. ^ r_ LO CM ■* 00 CT> T r— O , — cm *x> rv , ts J TJ- CM CM >— CM CM en co CM rs rj tfopo^tmvcoii- CJl ,_ O CM CM CM IV. O <— CTl CO U"> Iv. 0> CO 9 LO CM co CO CO CM CO CO i— I— *t CO r^ > ^ f-i-i-miNTf (Nt-i-M^iOMJi E C »> '-;OOlOcococO'— i/lNOBifli- O L) >— CMIv.'uOLOirirOTj-O^DcO'vi^pri m 0 CM<— CMCMIJICOrO^-i— ^"^i— CMCM *j 0.7 only. Complete list available from author. + Birds in their second year or older. ^ Recovery areas with < 0.7. ROUND-ROBIN MIGRATIONS Direct and indirect recoveries suggest there were several migration routes con- verging during preseason banding at Imperial Valley. The main migration route came down through Utah to the Imperial Valley. One segment of the population remained there for the winter, while others continued on. One of these routes out of Imperial Valley accounted for 16.0% of the male and 5.6% of the female band recoveries and was directed toward the Central and Mississippi Flyways (Table 12). Peak months on these two flyways for both direct and indirect re- coveries were November and December (Table 11). Thus, the birds were win- tering there. Approximately 78% were recovered in Texas and Louisiana. Another important segment of the population that passes through the Imperial Valley wintered in Mexico. Most of these pintails were recovered during No- vember, December and January in west coast states. Pintails that migrate to Mex- ico by way of California may return to the Canadian breeding grounds via the Gulf Coast and the Central and Mississippi flyways (Low 1949). This was called a counterclockwise round-robin migration. There is a possibility that some birds migrated to the Imperial Valley through northern California then returned to the Central Valley for winter. However, the more likely explanation is that the birds came through Utah to the Imperial Val- ley. From there they broke up into several segments, one of which turned north to the Central Valley. More of the indirect male recoveries from birds banded preseason in Imperial Valley came from the Central Valley (32.4%) than Impe- rial Valley (16.3%, Table 12). These birds probably arrived in the San Joaquin 152 CALIFORNIA FISH AND GAME Valley after passing through the Imperial Valley prior to the hunting season. Some remained there for the winter while others continued to the San Francisco Bay- Delta and Sacramento Valley. The route these birds took back to the breeding grounds was unknown. While the counterclockwise round-robin migration has been documented (Low 1949), this report documents a clockwise round-robin migration passing through Utah to the Imperial Valley then up to the Central Valley. Approximately 15% of the immature and 21% of the adult direct recoveries of pintails banded in Imperial Valley were recovered in northern California. Thus the clockwise round-robin migration is important in the California pintail harvest. If Los Banos banded pintails migrate through Utah then turn west to the San Joaquin Valley bypassing the Imperial Valley, more indirect recoveries would be expected from Utah than were recorded (males 1.1%-females 3.6%, Table 8). However, fall migration probably occurs through Utah and/or northern Califor- nia before the hunting season, because all of the preseason bandings at Los Banos were completed by the end of August. The scarcity of band recoveries from north and east of California for all California banding stations suggest an early mi- gration. Pintails wintering in Mexico, Central and South America and along the Gulf Coast by way of the Pacific Flyway, have farther to go than those wintering in the Central Valley, so may start their migration earlier. All 25 recoveries from coun- tries south of Mexico came from bandings in August during the 1950's and none in the 1960's and 1970's when banding was delayed until September. Honey Lake banded pintails have closer ties to the San Joaquin Valley and the Central and Mississippi flyways than do pintails from other banding stations in northern California. Consequently, some birds migrate through Honey Lake to the San Joaquin Valley and Mexico and/or the Central or Mississippi flyways in a counterclockwise migration. It is perplexing why pintails banded in August at Los Banos would be recovered in Alaska (2), British Columbia (5), Nebraska (4), and South Dakota (2) in Oc- tober of the same year. Are these examples of an early, rapid round-robin mi- gration, both clockwise and counterclockwise? As previously discussed, some Imperial Valley pintails follow a clockwise round-robin migration into the Central Valley. Those birds recovered in British Columbia and Alaska could be on a con- tinuation of this route. Baysinger and Bauer (1971 ) mention an instance of re- verse migration of an immature male pintail banded at Sacramento NWR on Au- gust 29, 1969, and recovered on October 11, in Alaska. Was this reverse migration or was it another instance of the clockwise round-robin migration? The six pintails recovered in Nebraska and South Dakota in October may have come from Alaska to California then over to the Central Flyway following a counter- clockwise migration. A pintail might winter in the San Joaquin one year and on the Gulf Coast the next and still be within its established migration route. Also, the Central Valley might be regarded as one large wintering area where a bird might winter in the San Joaquin Valley one year and the Sacramento Valley the next. The data sug- gest that this could have occurred and/or several sub-populations were banded at each station. However, recoveries obtained in this study suggest that most pintails returned to the same wintering area each year. MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 153 I assume that those pintails recovered in the north states of the Central and Mississippi flyways migrated from California to these states and then down the flyways. However, some pintails, especially those from the Imperial Valley, may take a more direct route to the Gulf Coast after spending the fall in Imperial Val- ley. In later years after being banded in California, some pintails may have by- passed California and migrated south in the other flyways. Some of the following examples could have resulted from reverse or round- robin migrations; an immature male banded at Mountain Meadows on August 16, 1955, and recovered in Idaho on October 11; an immature male banded at Honey Lake on August 15, 1952 and recovered in Alberta on October 13; and an adult female banded at Tule Lake NWR on August 31, 1951, and recovered that hunting season in Alberta. Twenty-seven direct recoveries were obtained in Brit- ish Columbia and nine in Alaska from banding stations in northern California. Nine direct recoveries were obtained in Utah from Imperial Valley banded pintails. One was an immature male banded September 17, 1954 and recovered 23 days later in Utah. These Utah recoveries could be examples of reverse mi- gration or pintails moving from the Imperial Valley to the north states of the Cen- tral Flyway with a stopover in Utah. CAUSES OF THE WIDE RANGING DISTRIBUTION The wide ranging migration of males may have resulted from males pairing with females from different breeding grounds. Such males would be recovered on migration routes different from which they were banded. Bellrose (1968) be- lieved that the wide dispersion of recoveries for some species is caused by the mixing of flocks on staging areas before migration. He stated that when flocks combine on staging areas, a small segment of a population may change flight routes and even flyways, thus accounting for a wider dispersal of bands than would seem warranted by the distribution of flight routes. The home range of breeding pintails on the nesting grounds is generally wider than other species of dabbling ducks and males are generally more mobile than females at all stages of the reproductive cycle (Derrickson 1978). These characteristics probably apply on the wintering grounds as well. Immatures, especially males, are more likely to be recovered at distant locations than adults; e.g., all 25 Central and South Amer- ican recoveries were those of immatures; 7 of 9 recoveries from the South Pa- cific, Japan and Korea were immatures. Questions and Conclusions Two questions remain unanswered: (i) Since the 1950's what has caused pintails to delay their migration to the Klamath Basin? The early migrants could now be overflying the Basin and flying directly to the Central Valley and beyond; (ii) What is causing the downward trend during the past 30 years in band re- coveries from Canada as indicated by both Klamath Basin and Gray Lodge bandings? This could have been caused by declining populations in Canada re- sulting in banding samples containing fewer pintails from Canada, thus declining proportions of recoveries. It also could be due to later opening dates of Canadian hunting seasons or earlier migrations resulting from poor nesting success. 1 54 CALIFORNIA FISH AND CAME MANAGEMENT IMPLICATIONS AND RECOMMENDATIONS Although much of this report deals with the confusing aspects of pintail mi- gration, e.g., clockwise and counterclock-clockwise migration, reverse migra- tions, direct recoveries from distance points, etc., they are of comparatively mi- nor significance within the overall pintail migration. The vast majority of pintails migrate directly from staging areas in southeastern Alberta, southwestern Saskatchewan and Alaska to the Central Valley where they spend the winter. All other routes of pintails banded in California are of less importance. However, the continued loss of good marsh and food sources in the Central Valley and Imperial Valley could lead to a loss of pintails from California to other wintering areas, e.g., Texas and Louisiana. Color marking or biotelemetry of pintails in the Central Valley during the fall and winter would improve knowledge of daily and seasonal movements within this valley and provide data on the size of subpopulation wintering areas. Because of early migrations of pintails prior to the hunting season, data on mi- gration routes from breeding grounds to California were scarce. Did most of the pintails banded on the coastal stations migrate down the coast from Alaska or were their routes the same as found for birds banded on the inland stations? A telemetry or color marking program on the breeding grounds would improve knowledge on preseason migration routes. A similar marking program in Imperial Valley also would be of value in better understanding round-robin migrations. Banding stations within the San Francisco Bay-Delta had a higher proportion of local recoveries than other stations in California. If the degree of homing to a particular wintering area could be used as an indication of the desirability to win- ter on that area, then it could be concluded that the San Francisco Bay-Delta, es- pecially the Suisun Marsh, was the favored wintering grounds for pintails. Un- fortunately, banding in this area was conducted only in the 1950's, and there have since been changes in distribution of recoveries in California. Thus, a leg banding program in Suisun Marsh, such as occurred in the 1950's would be helpful in de- termining what changes have taken place over the past 30 years. ACKNOWLEDGEMENTS Thanks is given to the staff of the Klamath Basin NWRs, Salton Sea NWR, and the California Wildlife Areas for cooperation in the banding programs and to the members of the Waterfowl Studies Project, California Department of Fish and Game, who did the trapping and banding. Thanks is also given to M. Miller, j. Bartonek, D. Gilmer, J. Nichols and P. Law for review of the manuscript. LITERATURE CITED Baysinger, E.B., and R.D. Bauer. 1971. A documented instance of reverse migration in the pintail. Auk, 88(2): 438. Bellrose, F.C, T.C. Scott, A.S. Haukins and S.B. Low. 1 961 . Sex ratios and age ratios in North American ducks. Illinois Natural History Survey Bull., 27(6): 391^74. 1968. Waterfowl migration corridors east of the Rocky Mountains in the United States. Illinois Natural History Biological Notes, 61, 24 p. . 1976. Ducks, geese and swans of North America. Stackpole Books, Harrisburg, PA, 544 p. MIGRATION AND DISTRIBUTION OF NORTHERN PINTAILS 155 Derrickson, S.R. 1978. The mobility of breeding pintails. Auk, 95(1): 104-114. Ceis, A.D., and F.C. Cooch. 1972. Distribution of the duck harvest in Canada and the United States. U.S. Fish and Wildl. Serv. Spec. Sci. Rep.— Wildl., 151, 11 p. Low, S.H. 1949. Migration of the pintail. In: Migration of some North American waterfowl. U.S. Fish and Wildl. Serv. Spec. Sci. Rep.— Wildl., 1: 13-16. Moisan, C, R.I. Smith, and R.K. Martinson. 1967. The green-winged teal: its distribution, migration and population dynamics. U.S. Fish and Wildl. Serv. Spec. Sci. Rep.— Wildl., 100, 247 p. Rienecker, W.C. 1965. A summary of band returns from lesser snow geese (Chen hyperborea) of the Pacific Flyway. Cal. Fish and Game, 51 (3): 132-146. . 1976. Distribution, harvest and survival of American wigeon banded in California. Cal. Fish and Game, 62(2): 141-153. -| 56 CALIFORNIA FISH AND CAME Calif. Fish and Came 73 ( 3 ): 1 56- 1 62 1 987 ANALYSIS OF THE DIETS OF MOUNTAIN SHEEP FROM THE SAN GABRIEL MOUNTAINS, CALIFORNIA12 WILLIAM M. PERRY3 Department of Biology California State University Northridge, CA 91330 JIM W. DOLE Department of Biology California State University Northridge, CA 91330 and STEPHEN A. HOLL San Bernardino National Forest Fontana, CA 92335 Abstract: Dietary composition and quality of mountain sheep in the San Gabriel Mountains were determined by microhistological analysis and crude protein deter- mination of fecal pellets. Five browse species together constitute 57% of the dietary makeup on an annual basis: mountain mahogany, Cercocarpus betuloides, the major shrub species taken, constitutes from 13.8 to 55.5% of the identifiable plant frag- ments in every month; California buckwheat, Eriogonum fasciculatum, and white sage, Salvia apiana, are important dietary components in winter and early spring; holly-leafed cherry, Prunus ilicifolia, is utilized most heavily in fall, but is a minor component in all months; silk-tassel bush, Garrya veatchii, is most frequently eaten in fall and winter. Grasses are eaten year-round, but are most important as dietary components (21.8 to 39.3%) from April through November. Forbs comprise only 2% of the identifiable food fragments in the diet. Crude protein content of the fecal pel- lets changes gradually from a high of 20% in April to a low of 9.9% in September. INTRODUCTION Of the populations of mountain sheep, Ovis canadensis nelsoni, currently in- habiting the mountain ranges of southern California (Weaver 1982), the largest (mean 660, 1976-1985) occurs in the San Gabriel Mountains, northeast of Los Angeles. In view of the obvious importance of dietary information in the man- agement of the population, the paucity of information on the food habits of the San Gabriel herd is surprising. Although dietary studies of mountain sheep in other locations are available, only three previous food habit investigations pertain to the sheep in the San Gabriel Mountains: Robinson and Cronemiller (1954) provided observational information on feeding preferences of the San Gabriel population on its summer range above 1900 m; Graham (1968) and Weaver et al. ( 1 972 ) reported on the food habits of the animals on their winter range, based primarily on direct observation. With the exception of an analysis of the stomach contents of a single sheep provided by Graham, none of these authors made any attempt to quantitatively assess the diet. 'Accepted for publication October 1986. 2Based on a thesis submitted in partial fulfillment of the requirement for the degree of Master of Science at California State University, Northridge. 3Current address: Botany Department, University of California, Davis CA 95616 DIETS OF MOUNTAIN SHEEP 157 The objective of this study was to identify seasonal variation in the compo- sition and quality of the diet of the mountain sheep in the San Gabriel Mountains. To accomplish these goals, we: ( i ) surveyed the vegetation to obtain a qualitative evaluation of forage availability; (ii) determined forage utilization by means of microhistological analysis of plant fragments in the feces; and (iii) assessed diets quality by determining crude protein content of the fecal material. STUDY AREA The study was conducted in the San Gabriel Mountains, Los Angeles County, California. This mountain range, extending approximately 100 km east to west, is bisected by numerous faults, and has many narrow ridgetops and steep eroding cliffs. Elevations range from 200 to 3000 m. Geologically, the mountains are rel- atively young; soils are rocky and poorly developed (Jaeger and Smith, 1966). Chaparral communities predominate on the lower slopes. At higher elevations there are coniferous forests; alpine communities occur on the highest peaks. Cli- mate is Meditteranean with hot, dry summers and cool, moist winters. Most pre- cipitation occurs from November to March. The study area is located just southwest of Mt. San Antonio and includes the lower half of Cattle Canyon and the adjacent south-facing slopes of Cow Canyon. The 810 ha area extends in elevation from 900 to 1500 m. Vegetation is primarily chaparral, part of which (Cow Canyon) burned in 1975. The study area, an im- portant winter range of the San Gabriel herd, supports about 1 30 mountain sheep (Holl and Bleich 1983), many of which are non-migratory animals. METHODS Five individual groups of fresh fecal pellets were collected each month. Pellets were collected immediately after defecation. Pellet groups were collected De- cember 1979, January and October through November 1980, and February through September 1 981 . All samples were from ewe bands, i.e. bands containing fewer than 25% rams; in addition to adult females, the bands sampled included lambs, yearlings of both sexes, and rams. Fecal pellet groups were air dried in paper bags and frozen in plastic bags. Be- fore analysis, pellets were oven-dried (60°C for 24 h) and weighed. Ten pellets (two from each group) were picked at random from the five monthly samples. These were ground in a Wiley mill through a 20 mesh (1 mm) screen to insure uniform fragment size. The material was then prepared following the procedure of Hansen et al. (1977), and soaked in 60% perchloric acid to separate the epi- dermis from underlying tissues. Equal amounts of the fecal samples for each month were mounted on five microscope slides in a modified gum syrup me- dium (Williams 1959). Plant fragments in the fecal material were identified and quantified as sug- gested by Sparks and Malechek (1968). On each of the five slides containing each monthly sample, 20 fields were selected randomly. Each of these 100 fields was observed at 100X with a phase-contrast microscope and the frequency of each identifiable plant species recorded. Shrub and forb fragments were iden- tified to species; all grasses were grouped together. From these data the percent relative frequency of each plant species for each slide was calculated. These val- ues were used to compute monthly and seasonal means, and standard deviations from the means, using frequency to density conversion as per Sparks and 158 CALIFORNIA FISH AND CAME Malechek (1968). (Since unidentifiable fragments found in some fields were in- cluded in the calculations, the relative percentages for identifiable species do not sum to 100.) Plant fragments were identified by comparison with a reference collection of 84 plant species collected from the study area. Epidermal peels of leaves, stems, flowers and fruits of each species were taken following the procedure of Mohan Ram and Nayyar (1977), preserved in FAA, and mounted on reference slides. In addition, reference slides were prepared of dried leaves ground and processed in the same manner as were fecal pellets. Photomicrographs of reference slides facilitated comparison with the fecal samples. A portion of each of the five fecal samples collected each month was analyzed for percent crude protein content using the micro-Kjeldahl method (AOAC 1960) by the Animal Nutritional Testing Laboratories, University of Nevada, Reno. RESULTS Browse species comprise 60% of the diet on a yearly basis. Browse is the ma- jor dietary component in every month, and in winter (January through March) it makes up more than 80% of the diet ( Figure 1 ) . Fourteen browse species were identified in fecal samples (Table 1). Of these, mountain mahogany, Cercocarpus betuloides, is the most common forage plant, with an annual per- cent frequency of 35%; its frequency ranges from 14% in September to over 55% in June. In every month but four ( March, April, August, September) it is the most common component of the fecal samples. Four other shrub species, Cal- ifornia buckwheat, Eriogonum fasciculatum, holly-leafed cherry Prunus ilicifolia, white sage, Salvia apiana, and silk-tassel bush, Garrya veatchii, occur with a fre- quency of from 5% to 7%. The nine other shrub species identified in the feces have a combined annual frequency value of about 3%. X Composition 100 Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Month of Year FIGURE 1. Percent composition of forage classes in the diet of mountain sheep by month. DIETS OF MOUNTAIN SHEEP 1 59 On an annual basis, grasses consistuted 22.5% in the diet. Grass utilization is highest from April through November and lowest from December through March (Table 1 ) . All forbs together comprise less than 2% of the identified frag- ments in the pellets. TABLE 1. Seasonal Means (± Standard deviation) of Fecal Composition Analysis. All Values Are Relative Percent Density (0 = No Individuals Detected). T = Trace (< 1%). Species Season Winter Spring Summer Fall Browse Cercocarpus betuloides 34.6 ±12.8 36.1 ±16.2 46.6: ±15.0 23.9±11.0 Salvia apiana 10.6: t9.1 10.8 ±13.6 0 0 Eriogonum fasciculatum 15.2: t8.9 12.3 ±12.7 T 0 Prunus ilicifolia 4.5 ± 5.9 2.1 ± 3.2 6.0 ± 4.6 11. 7 ±5.4 Garrya veatchii 9.4 ± 5.2 1.5± 2.1 2.6 ± 2.4 7.7 ±5.4 Heteromeles arbutifolia T T 1.5± 1.9 T Yucca whipplei T T T 1.7±1.9 Rhamnus crocea T 0 T 1.9±2.5 Encelia californica 0 T T T Ceanothus leucodermis 0 0 T T Cercocarpus ledifolius 0 0 T T Rhus ovata T 0 T T Rhamnus californica 0 0 T T Eriophyllum confertiflorum 0 0 0 5.6 ±5.9 Leptodacty/on californicum 0 T T 0 Forbs Viola purpurea 0 0 T 0 Castilleja affinis 0 0 T 0 Mentzelia laevicaulis 0 0 T 0 Grasses all species 9.8 ± 7.5 25.6 ±15.3 27.9 ±14.6 28.5 ±8.6 The occurrence of three of the five major browse species (California buck- wheat, white sage, silk-tassel bush) peaks from December through March, when grass utilization is minimal. California buckwheat and white sage are virtually ab- sent from the diet during the remainder of the year, but silk-tassel bush shows an- other peak of utilization during the fall, particularly the month of September. The fourth major browse species, holly-leafed cherry, is consumed mainly in fall (Ta- ble 1). Some of the minor components of the diet, both forbs and browse, vary sea- sonally in occurrence. Eriophyllum, Eriophyllum confertiflorum, is found in the diet only in September and November; its frequency in those months exceeds 8%. Toyon, Heteromeles arbutifolia, and redberry, Rhamnus crocea, are eaten most commonly in summer and fall, but are occasionally consumed in winter and spring. Chapparal yucca, Yucca whipplei, is utilized irregularly throughout the year in small quantities, but in no month is its percent frequency higher than 2.5% The maximum protein content of the fecal samples ( 20% ) occurs in April and the minimum (9.9%) in September. From April to September the protein con- tent drops gradually in each month, whereas from September to April it showed a fairly regular rise (Figure 2). DISCUSSION Our data indicate that for the mountain sheep population sampled, browse is the predominant dietary component. Grasses are the second most important 160 CALIFORNIA FISH AND CAME Jun Jul Month FIGURE 2. Mean percent crude protein (total nitrogen) in the fecal pellets by month (n = 5). Bars indicate standard deviations from the mean. constituent of the diet, but are eaten in only minor quantities in the winter months when what little grass available is dry. Forbs appear in the diet only rarely. Of the five species of browse that make up the bulk of the food items identified in the feces, mountain mahogany is by far the most common. This plant was es- timated to make up only about 6% of the shrub cover in the study site. In every month, however, it had a high frequency of occurrence in the fecal material, from 14 to 55%. It is unknown whether the plant is consumed so commonly be- cause it is "preferred" by the sheep or because the animals spend the bulk of their time on the steepest slopes (escape terrain) where this species is locally abundant. The four other browse species commonly eaten (California buck- wheat, holly-leafed cherry, white sage, silk-tassel bush) occur approximately as frequently in the fecal samples as would be expected from their abundance in the environment. All other shrub species identified in the feces appear to be of very minor importance in the diet. Four very common shrub species (chamise, Adenostema fasciculatum; yerba santa, Eriodictyon crass ifoli urn; manzanita, Arctostaphylos sp.; hoary-leafed ceanothus, Ceanothus crassifolius), together accounting for about 50% of the shrub cover in the study area, were never re- corded in the feces. Our findings are generally consistent with those reported by others on the diet of the sheep in the San Gabriel Mountains. Graham (1968) reported the stomach contents of a single sheep from the South Fork of Lytle Creek, about 16 km east of our study site and in the same vegetational zone, to be 65% grass and 25% buckwheat with occasional forbs, oak leafage, and acorns. The much higher per- centage of grasses than we report may be the result of "sampling error" (a single DIETS OF MOUNTAIN SHEEP 1 61 animal vs several in our study) or because of biases of different analytical tech- niques. In any case, the presence of large quantities of grass in the stomach is not inconsistent with our findings. Graham did not indicate the season of his sample, but the presence of acorns in the stomach suggests the animal died in the fall, a period when we found grass utilization to be high. Our findings also are generally consistent with the observations of Weaver et al. (1972) that mountain sheep in the Lytle Creek area feed mainly on browse but supplement this with grass and occasional forbs. Robinson and Cronemiller (1954), working at a higher eleva- tion ( > 1900 m) on summer ranges, also reported browse to be a preferred food with grasses eagerly sought in spring. The species of browse they reported being consumed differed from those we report, probably reflecting elevational differ- ences. Among the five most frequently utilized species these authors identified were curl-leaf mountain mahogany, Cercocarpus ledifolius, and two herbaceous species of buckwheat, Eriogonum ovalifolium and £ umbellatum. The diet of mountain sheep in other parts of its range often differs greatly from that of the San Gabriel herd. In Nevada, rumen contents were reported to av- erage more than 65% grasses, about 28% shrubs, and 6.5% forbs (McQuivey 1978). In Death Valley National Monument (Welles and Welles 1961 ) and in Ar- izona (Russo 1956), mountain sheep are reported to take browse from trees and shrubs more commonly than grasses and forbs. More recently, however, Ginnett and Douglas (1982) reported browse to be most important among the Death Valley sheep; in their study 55.5% of the forage was browse whereas grasses and forbs constituted only 38.4% and 2.7%, respectively; the proportions of each in the diet varied little with the seasons. On the other hand, Seegmiller and Ohmart (1981) found mountain sheep in the Arizona desert to have a diet of 54% browse, 31% forbs, and 8% grasses on an annual basis. Although the relative im- portance of each of these forage categories varied seasonally, browse was pre- ferred in summer, fall, and winter whereas forbs predominated in the diet in spring; grasses were taken in low quantities at all seasons. Although our data suggest that forbs are a very minor component of the diet, it is likely that our estimate of their occurrence is low. Delicate and fragile forbs do not survive the digestive process as well as grasses and shrubs (Free, Hansen and Sims 1970, Slater and Jones 1971, Jacobs 1973, Stewart 1967, Smith and Shandruk 1979), leading to their under-representation in fecal samples. This con- clusion is also supported by the visual observations of Graham (1968) and Weaver et al. (1972) that forbs are a significant portion of the diet of the San Gabriel herd. Although the utilization of mast crops in the San Gabriel Mountains has been documented (Holl and Bleich 1983, Graham 1968), we found no evidence of acorns or other fruits in the fecal pellets. Perhaps such materials do not survive the digestive process in an identifiable state. On several occasions sheep were seen eating the fruits of holly-leafed cherry; none was ever seen eating acorns. The gradual increase in fecal protein from fall to spring coincides with the southern California rainy season and with the concomitant emergence of annual grasses and the green-up of browse. This suggests that forage quality, specifically available protein, increases during this period and reaches a peak more or less co- incident with gestation and lactation, the time of greatest nutritional demand on the ewes (Maynard and Loosli 1969). Without additional data, however, we can not eliminate the possibility that other causes, such as the presence of high con- 162 CALIFORNIA FISH AND GAME centrations of secondary compounds (e.g. phenolics) and low digestibility (Mould and Robbins 1981, 1982), are responsible for the elevation of fecal ni- trogen output. ACKNOWLEDGMENTS We gratefully acknowledge the help of the following individuals in this study: M. Thor for aid in plant identification, B. Meyerhoff and T. Lynn for assistance in the dietary analysis, and J. Wehausen and V. Bleich for reviewing the manuscript and offering many helpful suggestions. This project was funded by the USDA, San Bernardino National Forest. LITERATURE CITED AOAC 1 960. Official methods of analysis of the Assn. Official Analytical Chemists, 9th ed. Assn. Official Analytical Chemists, Washington, D.C 1094 p. Free, J. C, R. M. Hansen and P. L. Sims. 1970. Estimating dryweights of foodplants in feces of herbivores. J. Range Manage. 23:300-302. Cinnett, T. F. and C. L. Douglas. 1 982. Food habits of feral burros and desert bighorn sheep in Death Valley National Monument. Desert Bighorn Council, Trans. 1982:81-87. Graham, H. 1968. Habitat studies in the San Gabriel Mountains bighorn range in California. Desert Bighorn Council, Trans. 1968:54-58. Hansen, R. M., T. M. Froppe, M. B. Gilbert, R. C. Clark and H. W. Reynolds. 1977. The microhistological analysis of feces as an estimator or herbivored dietary. Composition Analysis Lab. Report., Colo. State Univ., Fort Collins. 22 p. Holl, S. H. and V. C. Bleich. 1983. San Gabriel mountain sheep: biological and management considerations. USDA Forest Service Administrative report, San Bernardino National Forest. 136 p. Jacobs, J. 1973. A microtechnique index to pronghorn diet and sagebrush digestion coefficients. Wyo. Game Fish Dept., Rept. 49 p. Jaeger, E. C. and A. C. Smith. 1966. Introduction to the natural history of southern California. Univ. Calif. Press, Berkeley, Calif. 104 p. Maynard, L. A. and ). K. Loosli. 1969. Animal Nutrition. McGraw-Hill, New York. 613 p. McQuivey, R. P. 1978. The desert bighorn sheep of Nevada. Bio. Bull. No. 6. Nevada Dept. Fish Game. 81 p. Mohan Ram, H. Y. and V. C. Nayyar. 1977. A simple technique of obtaining epidermal peels from fresh, pickled, dried and processed leaves. Stain Tech. 52:53-54. Mould, E. D. and C. T. Robbins. 1981. Nitrogen metabolism in elk. ). Wildl. Manage. 43:323-334. 1982. Digestive capabilities of elk compared to white-tailed deer. ). Wildl. Manage. 46:22-29. Robinson, C. S. and F. P. Cronemiller. 1 954. Notes on the habitat of the desert bighorn in the San Gabriel Mountains of California. Calif. Fish Game 40:267-271. Russo, ). P. 1956. The desert bighorn sheep in Arizona. Ariz. Game Fish Dept., Fed. Aid in Wildl. Restoration Proj. Rept. W-44-R. 153 p. Seegmiller, R. F. and R. D. Ohmart. 1981. Ecological relationships of feral burros and desert bighorn sheep. Wildl. Mono. No. 78. 58 p. Slater, j. and R. |. Jones. 1971. Estimation of the diets selected by grazing animals from microscopic analysis of the faeces — a warning. Aust. Inst. Agr. Sci., J. 37:233-240. Smith, A. D. and L. J. Shandruk. 1979. Comparison of fecal, rumen and utilization methods for ascertaining Pronghorn diets. J. Range Manage. 34:275-279. Sparks, D. R. and K. C. Malechek. 1968. Estimating percentage dry weights in diets using a microscopic technique. |. Range Manage. 21:264-265. Stewart, D. R. M. 1967. Analysis of plant epidermis in faeces: a technique for studying the food preferences of graz- ing herbivores. J. Appl. Ecol. 4:83-111. Weaver, R. A., 1 982. Bighorn in California: a plan to determine the current status and trend. Calif. Dept. Fish Game, Fed. Aid in Wildl. Restoration Proj. Rept. W-51-R. 23 p. Weaver, R. A., J. L. Mensch, W. Timmerman and J. M. Hall. 1972. Bighorn sheep in the San Gabriel and San Bernardino Mountains. Calif. Dept. Fish Game, Wildl. Manage. Admins. Rept. 72-2:1-38. Welles, R. E. and F. B. Welles. 1961. The bighorn of Death Valley. U.S. Natl. Park Serv. Fauna. Series 6:1-242. Williams, O. 1959. Modified gum syrup. Turtox news 37:251. REFUTATION OF LENGTHS ATTRIBUTED TO THE WHITE SHARK 1 63 Calif. Fish and Game 73 ( 3 ) : 1 63-1 68 1 987 REFUTATION OF LENGTHS OF 11.3, 9.0, AND 6.4 M ATTRIBUTED TO THE WHITE SHARK, CARCHARODON CARCHARIAS" JOHN E. RANDALL Bernice P. Bishop Museum Honolulu, Hawaii 96817-0916 The 36.5-foot (11.1 m) length attributed to an Australian specimen of the white shark, Carcharodon carcharias, in 1870 is an error. The length determined from the jaws and teeth of this shark in the British Museum ( Natural History) is about 17.7 feet (5.4 m). Wood (1982) presented a photograph of a white shark taken in the Azores which he claimed was 29.5 feet (9 m) in length. He also attempted to lend authen- ticity to a report of a 37-foot (11.3 m) white shark from New Brunswick, Canada, re- corded by Vladykov and McKenzie (1935). Evidence is presented herein to refute both of these lengths. The 21 -foot (6.4 m) length given for a white shark caught off Cuba in 1945 is also erroneous. From a photograph of the shark, the 44-mm enamel height of an upper jaw tooth, and a vertebral centrum 80 mm in diameter, its length is estimated as about 16 feet (5m). The largest white shark reliably measured appears to be one from Ledge Point, Western Australia, which was 19 feet, 6 inches (6 m) long. Although it is probable that C. carcharias exceeds a length of 20 feet (6.1 m), irrefutable evidence of such a length has yet to be presented. Few marine animals have had as much written about them as the white shark, Carcharodon carcharias (Linnaeus), justifiably the most feared creature of the sea. Paradoxically, very little is known of the habits and biology of this shark. For example, there is no verifiable record of the capture of a female bearing young.2 We know nothing of the breeding habits, development, number of young in a lit- ter, etc. Also, not all that has been published on this species has been correct. Even something as straight-forward as the total length of specimens has at times been misrepresented in the literature. Randall (1973) examined the jaws of a white shark from Port Fairy, Victoria, Australia, in the British Museum (Natural History) (Figure 1) which Gunther ( 1 870) reported as 36% feet (11.1 m ) in length. Suspecting an error in the length, Randall measured the perimeter of the upper jaw and the vertical height of the enamel of the largest upper tooth (from the tip perpendicular to a line connecting the most ventral extension of enamel on either side). From comparable mea- surements of jaws and teeth of white sharks at various museums for which au- thentic total lengths had been obtained, he plotted two graphs: perimeter of the upper jaw against total length and vertical height of the enamel of the largest up- per tooth against total length. He then entered the measurements of the British Museum jaw and upper tooth into these graphs and concluded that the true total length of the Port Fairy shark approximated 17 feet, 8 inches (5.4 m). Perry W. Gilbert had also examined the jaws in the British Museum (Natural History) and suggested that there might have been a printer's error, 36% feet instead of 161/2 feet (Gilbert and Gilbert 1973). However, Gunther (1880) wrote that the white shark can attain 40 feet (12.2 m); evidently he rounded off the 36y2-feet. 1 Accepted for publication December 1986 2 See postscript 164 CALIFORNIA FISH AND CAME FIGURE 1 . Jaws of a white shark from Port Fairy, Victoria, Australia in the British Museum ( Natural History). The shark was reported by Gunther (1870) to have been 361/2 feet (11.1 m) in length, but comparison with other sets of jaws reveals the probable length at about 17 feet, 8 inches (5.4 m). On 19 April 1974, a letter was received from Gerald L. Wood of Guinness Su- perlatives asking why the 37-foot ( 1 1 .3 m ) white shark taken in New Brunswick in 1930 (Vladykov and McKenzie 1935) and other records above 21 feet had not been accepted as valid by Randall (1973). I replied that these lengths were ob- tained from fishermen without verification. No evidence in the form of jaws or teeth or photographs that accurately depict the length of these alleged enormous sharks was given. Mr. Wood was asked to pass on proof he might have of any REFUTATION OF LENGTHS ATTRIBUTED TO THE WHITE SHARK 1 65 C. carcharias in excess of 21 feet (6.4 m), the supposed authentic length given by Bigelow and Schroeder (1948) of a shark caught off Cuba in 1945. On 24 September 1980, Mr. Wood wrote that he had received a photograph of a white shark caught off San Miguel, Azores, in 1978 "which measured an as- tonishing 27 feet in length." He included a xerox copy of the photograph. The shark is foreshortened in the photograph, with people standing immediately be- hind it. It is difficult to judge the length of the shark from the copy, but from the size of the bystanders it does not seem as large as 27 feet. A letter was written to Mr. Wood asking if a print of the photograph might be sent and to inquire if the jaws or teeth were saved. Wood replied that there was an error in the length of the Azores shark; he wrote, "I have since learned that the actual length was 29 feet 6 inches or 9 metres." He added "I will not be able to let you have a print until May 1982, when the 3rd edition of my compilation, The Cuinness Book of Animal Facts and Feats, comes out. The teeth were purchased by a man living on Terceira, another island in the same group." Knowing the importance of examining the teeth of this shark and being unable at that time to visit the Azores, I contacted John E. McCosker of the Steinhart Aquarium to see if he might make such a trip. He, in turn, suggested this to Richard Ellis, the author of The Book of Sharks (1975). Ellis went to the Azores and made a major effort to find evidence of the landing of a white shark of ex- traordinary size. Of his visit he wrote, "I had been to three islands in the Azores and had called several others. I had been to the museum (where there was a stuffed 16-foot specimen), the newspaper offices, the harbors, the docks, the bars, the houses. I had drawn pictures of teeth, tried to ask for them in Portuguese and even offered a cash reward to anyone who could produce a big tooth. I came back empty-handed. . . ." (Ellis 1983). Further negative evidence was provided by McCosker who had an announcement printed in the Ponta Delagada daily newspaper requesting information on large white sharks. No responses were re- ceived. The third edition of the The Guinness Book of Animal Facts and Feats was pub- lished in 1982, and the photograph of the alleged 29.5-foot white shark from the Azores appears on pages 132-133 (reproduced herein as Figure 2). Attention is drawn to the legs and hands of the people standing just behind the shark, relative to the maximum height of the shark. Figure 3 is a photograph of a 16-foot 7-inch (5.1 m) white shark caught off Okinawa. Compare the height of this shark rel- ative to the men standing beside it. It is obvious that the Azores shark could not have been 29/^ feet long — indeed probably not over 20 feet — since the Okinawa shark is 16-feet 7-inches long. Wood (1982) discussed his attempt to authenticate the 37-foot length of a white shark caught in a herring weir at White Head Island, New Brunswick in June, 1930, which was reported by Vladykov and McKenzie (1935). He suc- ceeded in contacting Mr. R. A. McKenzie who stated that Dr. Vadim Vladykov had supplied this record, but added that the shark had not been examined by him; his information had been secondhand. Wood wrote, "Would the Canadian ichthyologist [Vladykov], a man of high reputation, have included heresay data unless he was positive it was accurate?" Wood's attempts to contact Vladykov "proved fruitless." I also tried to communicate with Vladykov (who has since 166 CALIFORNIA FISH AND CAME FIGURE 2. White shark caught off the Azores reported as 291/2 feet (9 m) in length by Wood (1982). M *' 4 INK FIGURE 3. A 16-foot, 7-inch (5.1 m) white shark caught off Okinawa (Kyodo Newsservice Photo). REFUTATION OF LENGTHS ATTRIBUTED TO THE WHITE SHARK 1 67 died), but failed; however, Dr. Don E. McAllister, National Museum of Natural Sciences, Ottawa, Canada, did so on my behalf and wrote (11 July 1984), "Vadim remembered the record. He did not see the specimen and said the fish- erman had provided the length measurement. He did recall that the fisherman had given him two teeth and that he still had them." Only one tooth was supplied which Vladykov said was from the White Head Island specimen. On the back of the tooth is written "1936" which is not the date of collection although it is the date of publication. A xerox copy of the tooth was sent to me, along with its enamel height of 28 mm. From the size and shape of the tooth and its height rel- ative to its width (width of enamel at base 1.2 in height of enamel in midline), one can ascertain that it is from the front of the lower jaw. Comparing this with lower front teeth in white shark jaws in the collection of the California Academy of Sciences, I conclude that it came from a shark of about 16 to 17 feet in length (5.0-5.3 m). Therefore we again return to the 21 -foot length given by Bigelow and Schroeder (1948) for the shark caught off Cojimar, Cuba, in 1945 as the largest reported in the scientific literature. However, even this length is incorrect. Guitart and Milera (1974) illustrated this shark with numerous people standing just be- hind it; one man is leaning over with the hands of his outstretched arms on the back of the shark just in front of the first dorsal fin. If one compares the relative height of these people and the Cuban shark to those standing by the Okinawa shark, the latter shark seems larger. An upper jaw tooth from the Cuban shark (presumably a large one from the front of the jaw) is figured beside a millimeter scale by Guitart and Milera; from this an enamel height of 44 mm may be de- termined. Such a tooth size is found in a shark of about 16 feet (4.9 m) in length. A vertebral centrum from the shark was also illustrated; it was measured as 80 mm in diameter and 37 mm in width. If this centrum was taken from under or an- terior to the first dorsal fin, the regression of vertebral centra given in Cailliet et al. (1985) would indicate a shark length of 16 feet 4 inches (5 m). The weight of the liver of this shark was given as 1,005 pounds (456 kg). The total weight of the shark was estimated by fishermen at 7,000 pounds (3,175 kg). A shark of this size, however, would have had a liver much heavier than 1,005 pounds (John Hewitt, Steinhart Aquarium, pers. comm.). Curiously, the point representing the 21 -foot length and 7,000-pound weight of the Cuban shark fits well at the upper end of the length-weight curve for C. carcharias given by Tricas and McCosker (1984:224). It seems that the exag- gerated length and the estimated weight were of the same order of magnitude. Bigelow and Schroeder (1948) reported the second largest white shark (after the "21 -footer") "actually measured" was 19 feet in length, but they gave no lo- cality or other details on the shark. They added, "We should perhaps caution the reader that estimates of the size of the larger sharks are frequently much too high; e.g., an Australian specimen, reported in the local newspapers as 16 feet long, ac- tually measured only eight feet six inches." In recent years Dr. Gordon Hubbell of Key Biscayne, Florida has been keeping records of the length of large white sharks, the perimeter of their upper jaws and the perpendicular height of the enamel of the largest upper teeth. His largest record is a 19-foot 6-inch (6 m) female caught at Ledge Point, Western Australia on 22 March 1 984. The perimeter of its upper jaw is 1 30 cm and the height of the enamel of the largest upper tooth 51 mm. 168 CALIFORNIA FISH AND CAME Undoubtedly Carcharodon carcharias exceeds 20 feet (6.1 m) in length, but as yet there is no authenticated record of such a size. If a large white shark is captured, particularly if it is larger than 1 9 feet, 6 inches, every effort should be made to properly document its size. A horizontal mea- surement should be carefully taken from the tip of the snout to the end of the up- per lobe of the caudal fin. This is the straight-line distance, not the curved mea- surement over the surface of the body. The girth might also be measured. A photograph should be taken with the shark perpendicular to the direction of the camera and something such as a meter stick or yard stick set on the back of the shark. The jaws with the teeth intact and some vertebrae should be saved and de- posited in a major museum. The height of the enamel of the largest upper-jaw tooth and the perimeter of the upper jaw should be measured. The vertebrae should be taken from beneath or anterior to the first dorsal fin. If possible, the to- tal weight of the shark should be accurately measured before it is cut open. It would be of interest to obtain the weight of the liver separately. Also any prey animals in the stomach should be weighed, and their weight subtracted from the total. If one wishes to make the many different measurements of the body and fins of a shark, see the diagrams of Compagno (1984). ACKNOWLEDGMENTS I am grateful to the following persons for providing valuable information for this article: Richard Ellis, Robert T. B. Iversen, John Hewitt, Gordon Hubbell, Don E. McAllister, John E. McCosker, Timothy C. Tricas, and Gerald L. Wood. Mc- Cosker and Tricas reviewed the manuscript. LITERATURE CITED Bigelow, H.B., and W.C Schroeder. 1948. Sharks. In, Fishes of the western North Atlantic, no. 1:59-576. Sears Foun- dation for Marine Research, Yale University, New Haven. Cailliet, CM., L.J. Natanson, G.A. Welden, and D.A. Ebert. 1985. Preliminary studies of the age and growth of the white shark, Carcharodon carcharias, using vertebral bands. Symposium on the biology of the white shark. Mem. S. Calif. Acad. Sci., 9:61-72. Compagno, L. 1984. FAO species catalogue, vol. 4, part 1, Sharks of the World, 249 pp., Food and Agriculture Or- ganization of the United Nations, Rome. Ellis, R. 1975. The book of sharks. Grosset & Dunlap, New York. 320 p. 1983. Chiller from the depths. Ceo, March (19831:92-97. Gilbert, P.W. and C. Gilbert. 1973. Sharks and shark deterrents. Underw. Jour., 5(2):69-79. Guitart, D. and J.F. Milera. 1974. El monstruo marino de Cojimar. Mar y Pesca, 104:10-1 1 . Giinther, A. 1870. Catalogue of the fishes in the British Museum, vol. 8. Taylor and Francis, London, xxv + 549 p. 1880. An introduction to the study of fishes. Adam and Charles Black, Edinburgh. 720 p. Randall, J. E. 1973. Size of the great white shark (Carcharodon). Science, 181:169-170. Tricas, T.C. and J.E. McCosker. 1984. Predatory behavior of the white shark (Carcharodon carcharias), with notes on its biology. Proc. Calif. Acad. Sci., 43(14):221-238. Vladykov, V.D. and R.A. McKenzie. 1935. The marine fishes of Nova Scotia. Proc. Nova Scotian Inst. Sci., 19(1):1-113. Wood, G.L. 1982. The Guinness book of animal facts & feats, ed. 3. Guinness Superlatives Ltd., Middlesex. 252 p. Postscript — In a symposium entitled "The Systematics, Ecology and Physiology of Elasmobranchs" held at the University of Tokyo, 17-18 March 1987, S. Uchida of the Okinawa Expo Aquarium re- ported on the capture on 2 April 1986 of a female of C. carcharias off Wakayama, Japan which was about 470 cm total length. It contained seven nearly full-term embryos about 100-11 0 cm total length. Unfortunately, neither the mother shark nor the embryos were retained. Uchida has prepared a manuscript on this discovery which will feature a photograph taken by I. Wakabayashi of the female and its embryos. REPRODUCTIVE RHYTHMICITY OF THE ATHERINID FISH 1 69 Calif. Fish and Game 73 ( 3 ): 1 69- 1 74 1 987 REPRODUCTIVE RHYTHMICITY OF THE ATHERINID FISH, COLPICHTHYS REGIS, FROM ESTERO DEL SOLDADO, SONORA, MEXICO1 2 G.A. RUSSELL 194 Yale Avenue Winnipeg, Manitoba, Canada, R3MOL8 D.P. MIDDAUGH AND M.J. HEMMER U.S. Environmental Protection Agency Environmental Research Laboratory Gulf Breeze, Florida 32561 ABSTRACT The reproductive rhythmicity of the atherinid fish, Colpichthys [c.f. Atherinops] regis (Jenkins and Evermann 1889), was observed in the Estero del Soldado, Sonora, Mexico during October 1982 through January 1983. Spawning runs occurred at ap- proximately two week intervals during daytime high tides. These high tides coin- cided with new and full moons. Spawning only occurred when predicted tidal heights were >0.73 m above MLW. Eggs were deposited in the upper intertidal zone in locations that appeared to provide protection from predators, thermal stress and desiccation. INTRODUCTION Fortnightly reproductive rhythms have been identified in several intertidally spawning atherinid fishes. The best known of these is the California grunion, Leuresthes tenuis, which is found along the Pacific Coast from Monterey Bay, California to Bahia Magdalena, Baja California Sur (Moffatt and Thomson 1978). Spawning runs of L tenuis occur from February through August (Clark 1925; Walker 1952). Females deposit eggs in sand on nighttime high tides, just after the highest high tides during each semi-lunar period (Shepard and LaFond 1940; Middaugh, Kohl and Burnett 1983). The Gulf grunion, Leuresthes sardina, also has a fortnightly spawning periodicity and deposits its eggs in sand. Spawning oc- curs from January to May in the northern Gulf of California. Due to a shift in the time of high tides that occurs between January and May, the Gulf grunion spawns during nighttime in the early part of the reproductive season and in daytime to- wards the end of the season (Thomson and Muench 1976). The Atlantic silverside, Menidia menidia, with a range from the Magdalen Is- lands, Quebec, Canada to northern Florida (Cox 1921; Johnson 1975), is a rhyth- mic spawner that deposits eggs on vegetation in the upper intertidal zone during daytime high tides from March through July (Middaugh, Scott and Dean 1981 ). Maximum intensity spawning runs occur on days just after new and full moons (Middaugh 1981). The precise role of environmental variables in the observed spawning rhythms in these atherinid fishes is not fully understood, nevertheless, each exhibits a "lu- nar periodicity". Our study of the false grunion, Colpichthys [c.f. Atherinops] Accepted for publication February 1987. : Contribution No. 586 of the Gulf Breeze Environmental Research Laboratory and No. 670 of the Belle W. Baruch Institute, University of South Carolina, Columbia, S.C 29208. 1 70 CALIFORNIA FISH AND CAME regis (Jenkins and Evermann 1889)*, an atherinid fish occurring in the northern Gulf of California, was conducted to determine if this species also demonstrates a reproductive rhythmicity and to elucidate the potential influence of environ- mental variables on the timing and location of spawning runs within a small es- tuary ("estero") on the coast of Sonora, Mexico. MATERIALS AND METHODS Study Area Observations were made at a site known locally as Estero del Soldado ( Lat. 27° 57' 31 " N, Long. 1 1 0° 59' 00" W ) , located between the city and port of Guaymas, and San Carlos, a small community ca. 19 km WNW of Guaymas, Sonora, Mexico. The estero is in a crescent-shaped ( horns pointing south ) low-lying area whose northern boundary is an interrupted arc of hills and low mountains that extends from slightly west of San Carlos to the eastern side of the estero. Two parallel and interdigitating zones of low-growth mangrove trees form a narrow fringing forest (Lugo and Snedaker 1974) around the estero from south to west. At the water's edge, the zone is composed of red mangroves, Rhizophora mangle. A band of black mangroves, Avicennia germinans, is gen- erally located landward of the red mangroves. Shoreline substrate is composed of reddish-brown marine clays with abundant bryozoan remains and Chione shells backed by sand. As water depth increases, the substrate changes to brownish-black, foul-smelling silts and sands. Observations on the reproductive habits of C. regis were made in the upper reaches of the westward most arm of the estero at two stations; the first in a small stand of red mangroves, the second in a nearby man-made pile of tile rubble slightly higher in the intertidal zone at the base of a small stand of black man- groves. Environmental Measurements All environmental measurements were taken daily between 0700 and 0900 hrs. During October 1982 through March 1983, salinity was measured to the nearest 0.2 %o for samples collected at station 1 (Strickland and Parsons 1968). Water temperature was determined with an immersion thermometer and dissolved ox- ygen was measured by the modified Winkler Procedure (American Public Health Association 1975). Tidal predictions (times and heights in meters for Guaymas) used in our anal- yses were taken from Universidad Nacional Autonoma de Mexico (UNAM), Instituto de Geofisica, "Tables de Prediccion de Mareas, Puertos del Oceano Pacifico" (1982, 1983). Biological Observations The reproductive periodicity of C. regis was determined by one of us (GAR) watching the spawning sites daily (0600-1100 h) from 18 October 1982 through 9 March 1983. Synoptic observations during October through April of 1979-1981 revealed that C. regis only spawned in the upper intertidal zone on 'Note on the genus taxonomy. The original describers of the species placed it in the genus Atherinops [)enkins, O.P.,and B.W. Evermann 1889., Proc. U.S. Natl. Mus., Vol. 1 1 (for 1 888) :1 38-1 39). C.L. Hubbs [Proc. Acad. Nat. Sci. Phila., 1918, Vol. 69 (for Oct. — Nov. 1 91 7) :67] thought the species distinctive enough to erect a new genus (Colpichthys) to contain it. Recent evidence on the biochemical systematics of the atherinids suggests that regis may belong in the genus Atherinops. REPRODUCTIVE RHYTHMICITY OF THE ATHERINID FISH 171 morning high tides. No reproductive activity was observed at other times or lo- cations within the area of the stations. Because it was difficult to accurately assess the intensity of each run ( number of fish involved ) , we scored daily observations as 0 — no run observed; or 1 — spawning observed. This conservative approach was used because Binkley (1973) has shown that loss of amplitude (in this case, the number of fish in a spawning run ) has little effect on time series observations. When a spawning run was observed, the time that the run began was noted. The time when a run ended was also recorded on several occasions. For ease in anal- ysis and reporting, clock times were converted to decimal hours (i.e. 0630 = 06.50 h). All statistical analyses were conducted using programs pro- vided in the SAS Institute Inc. User's Guide (1982). RESULTS AND DISCUSSION Reproductive periodicity. — Spawning runs in Colpichthys regis occurred on daytime high tides between 07.08 and 10.13 hrs. The runs coincided with the fort- nightly occurrence of tides of maximum height during the approximate time of new and full moons (Figure 1 ). During spawning runs, the mean water temper- ature was 13.7°C (range 10.5-16.5); salinity 36.2%o (35.4-36.9) and dissolved oxygen 5.6 mg/L (4.1-6.0). I.O-i 5 0.8- I- i UJ I t o o o UJ or o. 0.6- 0.4- 0.2- 0.04 15 OCT 1 982 — i — 15 NOV 30 — i — 15 DEC 31 — i — 15 JAN 15 FEB 1983 TIME FIGURE 1 . Reproductive periodicity in Colpichthys regis. Values above each peak in spawning runs indicate the number of days runs occurred. Predicted tidal heights (m) are for daytime high tides. Open circles = full moons, closed circles = new moons. After observing several spawning runs, it became apparent that at least two fac- tors influenced spawning by C. regis. The first was occurrence of tides of suffi- cient height to make the spawning substrates available. Spawning only occurred 1 72 CALIFORNIA FISH AND GAME on days when predicted tidal heights were > 0.73 m above mean low water (MLW) and when these high tides occurred between 07.08 and 10.13 h (Table 1 ). Secondly, an increase was noted in the delay between the predicted (for Cuaymas) and observed (for Estero del Soldado) times of high tide with in- creased tidal heights (Table 1 ). A non-parametric Spearman rank correlation test was conducted to determine if a correlation existed between the predicted tidal height vs. the delay between predicted and observed times of high tide. There was a significant positive correlation (r=0.63, P = 0.015) between tidal height and delay between predicted and observed times of high tide. We attribute this significant correlation to hydrological conditions at the mouth of the estero. Wa- ter must flood and ebb from the estero through a restricted channel that is only 76 m wide at the highest tide and 15 m wide at low tide. In contrast, the total sur- face area of the estero is 1 .43 km 2. Tides at the estero entrance continue to flood for at least one hour after the predicted time of high tide for Guaymas and may ebb for up to two hours after the predicted time of low tide (UNAM, Instituto de Geofisica "Tablas de Prediccion de Mareas . . . ," 1982, 1983). A paired comparison T-test for differences between predicted and observed times of high tide revealed a significant difference (P = 0.01 ) for the two vari- ables. However, a check of Table 1 shows that the observed times of high tide in the estero and observed onset of spawning runs are nearly identical. A paired comparison T-test of observed times for high tides and the time of spawning runs showed no significant difference (P=0.15). The precise timing of spawning runs to coincide with high tide has been noted in other intertidally spawning atherinid fishes. Atlantic silversides, Menidia menidia, from the North Edisto River estuary in South Carolina spawn within a narrow range of intertidal elevations (1.4 to 2.4 m above MLW) and depend on tides of sufficient height to inundate intertidal spawning substrates (Middaugh and Takita 1983). In contrast, M. menidia, in Salem Harbor, Massachusetts gen- erally spawned during ebb tides with spawning activity occasionally continuing for up to 2.5 hours after high tide. Spawning usually occurred in water 1 cm or less in depth (Conover and Kynard 1984). Colpichthys regis always spawned in very shallow water. Indeed, individuals spawned at station 1 as long as their bodies could be exposed to the atmosphere while egg deposition and fertilization occurred. As the water level rose, spawning fish sometimes moved to station 2 (rubble pile) which was slightly higher in the intertidal zone. This behavior indicates that C. regis spawns as high as possible in the intertidal zone. The 3.0 mm diameter eggs were deposited in succulent vegetation growing at the base of black mangroves. Eggs deposited over the pile of loose rubble and broken ceramic tiles generally sank into cracks and crevices where they were apparently protected from desiccation and thermal stress. On one occasion C. regis was observed spawning in the entrance of an ap- parently abandoned crab burrow at the water's edge. As the tide receded, fer- tilized eggs were found 2 to 3 cm deep in the crab burrow. The high tide on this day failed to reach the usual spawning substrates. It seems likely that the crab bur- rows would provide some protection from environmental stresses. Indeed, Middaugh et al. (1981 ) observed that another atherinid fish, M. menidia, utilized abandoned crab burrows along erosional scarps as a spawning substrate. m q. <*> _ 1 ! Ol 0 1 ■ft 0 E REPRODUCTIVE RHYTHMICITY OF THE ATHERINID FISH 1 73 u ■o l/> -D O T^ ^ , 8? 8S , S rg O U (/) 2 "5 "S*o 01 o O £ "O LU 2 -S5 = .s». s o 4§ Oj 0 r o o3< ^dr-'r-'dd^ddddddci Q +++++++ I + I + I + IS "£* ro CO (N i <- ; O CO « ., ocrio*io*iodo6od(v!(v!t^rs.h-!odcTi w •* £ «J — o o o o o o o o o o o o o 0| !5 •- .C "O DC V ■s i o .o u o E-o 1 a> 'Z. c ■a ■o v c -. (0 CI us » £ »> *2 *" 0) ri" "$>■£ e C ifl — in ec ev • — k- CV nj f 5 2 4* 2 "3 *" c ■o o -r-O(i-f»imN(»lNe0C^ --~ ^» -».-«.■ . ^~^.,— ,_ (-M <-v| ^T<^ ^T^> 7^0^ ^TOO ^rr-^ tJ- ^ro ^rs ' C^S vO I ' I tN I «— ■ T— It— I P $ TNlir-COr-VON^T- Kr-Or-lfl ' t! fK 2 *£ ' ' IfNI I r— Ir— c LO rf ■5,fe Jl Q *- oo 4 i- rv o ~ — ft. ■— ' ' — ' ' — ' ' — ' ' — ' ' — ' II Is O £J5 I T— I I CO I r- I r- I I s s —<*: o2 ™tz ^ "^ r> «S o^ s^ O- ^PrlWi'- •5C ",' 00 (s " r^ O-l r- (S m < <0 _j ^ to to Q op rr — ' 5 fc ?^5T^"^So^?7 Q. ** c (N I— m*£ -; o> ■ •— in ^D i— t\ ,— rs "™ 4 vi "5 Jr. = *; lv o^ riroo ■ — 'N jriC—cor-in jrvO^ - 00 — ■ »— • rsj . o> -CO i£ 0> TTTfN — m E«f- | — 'O — O o> rr,(N ^^m ^t^t— ' — '^o ^^m ^^m ' — i -^ i» rf w. vo . m • t *_• «— . N kj >- <- ' — ' 0> l-v (N ' — ' >~ a. o a. .83 |^ oo 31~6*si^in J> O ^ C — TT T- *— O A O >— 7— ,0>rr,0>'7r">Drr,fn'— 't' s 01 c < 3 c « -c a. U c u *~ '= F — V 0) > \\ -c ?! a. *• 0 i 5c2 "J o> -5 1 — ' o> o> I £ i2T = at Q < 5 0 00 -C § tn 3 s 0 QJ M >- c n u E c COMPARISON OF SACRAMENTO SUCKER 181 75- 50 25. Sacramento 15 9 6 10 2 6 0 Russian 14 11 2 8 0 9 0 Goose 3 11 7 2 0 14 3 Eel 13 8 4 2 4 11 2 Mad 7 2 0 3 1 4 0 Pajaro 1 1 0 2 11 1 21 Salinas 6 0 0 0 11 5 15 FIGURE 2. Dendrogram of cluster analysis using Ward's method and Euclidean distance to group Catostomus occidentalis specimens based on meristic characters. Separation distance is the percent of the Euclidean distance between least-alike specimens. Number of spec- imens per grouping from each population is listed. 75. 50- w 25. Sacramen o 12 4 0 0 3 4 10 3 3 Russian 20 8 0 0 1 4 14 0 3 Goose 19 7 0 0 3 1 2 1 1 Eel 18 2 3 0 1 0 8 4 2 Mad 5 2 1 0 0 1 6 1 0 Pajaro 21 0 5 6 0 0 0 1 0 Salinas 16 7 3 5 0 0 5 0 0 FIGURE 3. Dendrogram of cluster analysis using single linkage and Euclidean distance to group Catostomus occidentalis specimens based on meristic characters. Separation distance is the percent of the Euclidean distance between least-alike specimens. Number of spec- imens per grouping from each population is listed. 182 CALIFORNIA FISH AND GAME No differences were apparent among Goose Lake and Sacramento, Russian, Eel and Mad River populations for any scale counts compared (Figures 4-7). There was also little difference between Pajaro and Salinas populations. How- ever, Pajaro and Salinas River populations differed from other populations in their values for LLS ( Figure 4 ),SALL (Figure 5), and SRBD (Figure 7). There was slight overlap in LLS and SALL counts. No differences were shown in SBLL counts ( Fig- ure 6). N Sacramento 50 50 40 46 17 I BH 1 Russian rin Goose Eel r-*-i Mad Pajaro 40 rfri Salinas 40 r nin 60 65 70 ■ ■ i ■ i 75 Counts FIGURE 4. Variation in lateral line scale counts between Sacramento sucker populations. Included are range (horizontal line), mean (vertical line), one standard deviation on either side of the mean (light horizontal bars), and two standard errors on either side of the mean (dark horizontal bars). N Sacramento 50 Russian 50 Goose 40 Eel 46 ndn Mad 17 Pajaro 40 Salinas 40 i±L I 1 1 1 1 1 1 11 12 13 14 15 16 17 Counts FIGURE 5. Variation in scales above lateral line counts between Sacramento sucker populations. COMPARISON OF SACRAMENTO SUCKER 183 Sacramento 50 Russian 50 Goose 40 Eel 46 Mad 17 Pajaro 40 Salinas 40 10 11 Counts 12 FIGURE 6. Variation in scales below lateral line counts between Sacramento sucker populations. Sacramer ito 50 Russian 50 Goose 40 Eel 46 Mad 17 Pajaro 40 Salinas 40 .nfcL 25 30 Counts — i — 35 FIGURE 7. Variation in scale rows before the dorsal counts between Sacramento sucker populations. Based on comparisons of morphometric characters, discriminant analysis grouped most Sacramento sucker specimens into their actual populations (Table 3). Slight overlap was observed between Goose Lake, Sacramento River, and Eel River populations. A number of Mad River, Pajaro River and Salinas River spec- imens were also misclassified. Eye diameter, SNL and ABL values were strongly correlated with SL for most populations (Table 2) so were not included in discriminant analysis. 184 CALIFORNIA FISH AND GAME ,3$ o Z- P •- ^■i (£<* iA k i) o , 5 c < J5 c 3 c O 0. *E ■a 1> „ ■5 $ (J U 5 Cfc .2 T? ^ — — T u 3 ^J C Qi Li 3 E u C UJ -J X# .5 3 a; O c CO to c* (J < COMPARISON OF SACRAMENTO SUCKER 1 85 DISCUSSION Although specific results between the two clustering techniques varied, overall findings were similar; neither technique distinguished among populations of C. occidentalis. This disagrees with earlier findings (Snyder 1908, 1913, Fowler 1913), but supports recent authors questioning of subspecific status given to dif- ferent Sacramento sucker populations (Moyle 1976, Lee et al. 1980). Analysis with Ward's method finds tight "minimum variance" clusters, whereas analysis with single linkage finds "natural" clusters which are not necessarily tight. When the two results coincide, a reliable classification has been obtained (Wishart 1979). Results of other analyses varied with methodology. The graphic method of Hubbs and Hubbs (1953) showed no differences in scale counts among Sacra- mento sucker populations other than those from the Pajaro and Salinas Rivers. However, discriminant analysis of morphometric characters showed significant differences even within most subspecies (Table 3), but those specimens not clas- sified into their true population were not necessarily classified with specimens from the other population of the same subspecies. Synder (1908) did not report differences between Russian and Sacramento basin populations, but discriminant analysis comparing morphometric character values from his study showed these populations to differ significantly. Snyder ( 1 908) also reported no differences be- tween Eel and Mad River populations. Discriminant analysis distinctly separated Eel and Mad River populations. Sacramento sucker populations exhibit large variability in many of their morphometric and meristic characters. Rutter (1908) noted that specimens from certain localities have distinctly large or small heads. Specimens studied and compared here showed a wide range of HL's within each population (Table 2). Rutter also found that LLS counts ranged from 60-84 in suckers from the Sacramento-San Joaquin basin. This study found a range in LLS counts from 62-76 in fishes of just the Sacramento River basin. Snyder (1908) noted that "dif- ferences between the mouth and eye will usually serve to distinguish Humboldt specimens from Sacramento specimens at a glance." Rutter (1908) found that mouth and lip size vary greatly between specimens from different locations in the Sacramento basin. This great variability would seem to provide some difficulty in actually separating Humboldt and Sacramento specimens at a glance. Rutter (1908) also stated that there is no relation between lip size and the portion of the Sacramento-San Joaquin basin, i.e., there is no clinal variation due to location: There is no relationship between any particular variation and the division of the basin in which the specimens were taken. . . . Taken as a whole, the spe- cies is extremely variable. Specimens from one location often have a distinct physiogomy and can, as a whole, be readily distinguished. Pleasant Valley specimens are remarkable for big lips and fine scales, while North Yuba specimens have big lips and coarse scales. The Wolf Creek specimens have small heads, and so on. Subspecific designation of Sacramento sucker populations requires consis- tency in methodology and criteria. Sacramento sucker populations have shown so much variation that consistency would demand designation of numerous subspecies if based on these slight morphometric and (or) meristic variations, and on some degree of geographic isolation. For example, populations in streams 186 CALIFORNIA FISH AND GAME tributary to Tomales Bay show a slight difference in scale counts and morphometric characters from populations in the nearby Russian River system (Snyder 1914). Sacramento suckers from Tomales Bay tributaries could then be considered a separate subspecies depending on the criteria used. Upper Pit River specimens have also been found to have distinct scale and dorsal fin-ray counts (Martin 1967). Martin (1967) noted that subspecific designation of Pit River populations creates a problem, and that consistency requires placement of all populations from a distinct geographic area under one name. He found that some Pit River populations were not distinct, whereas others were. This is a typical oc- curence with Sacramento sucker populations (Rutter 1908). It is easier to be more consistent if variations within and among Sacramento sucker populations are considered as only variations within the species. If subspecific designations are to be used it would be necessary to examine and document variations among and within populations from virtually every stream and lake within the Sacra- mento sucker's range. The findings of this study have shown that most Sacramento sucker popula- tions do not exhibit subspecific differences. However, the Pajaro and Salinas River populations are distinctive in that they exhibit scale count differences ex- ceeding those recommended by Hubbs and Perlmutter (1942) for subspecific separation. Although these findings are inconsistent with those from cluster anal- ysis, Pajaro suckers may meet the subspecific definition of Mayr (1969) because they are the only populations that seem to be taxonomically distinct. Further study of the Pajaro and Salinas River populations is needed to determine their sta- tus. We recommend an electropheretic analysis to compare Pajaro and Salinas River populations to other populations for a final diagnosis. Prior to such a di- agnosis, we feel it best to use the geographic vernacular designation (Wilson and Brown 1953) of Pajaro sucker to distinguish these populations. High variability among and within other populations precludes designating them as being subspecifically distinct. ACKNOWLEDGMENTS We thank the staff of the California Academy of Sciences Ichthyological Col- lection and Sacramento State University Vertebrate Collection for their cooper- ation and encouragement. We would also like to acknowledge E. M. "Morgan" Dixie for his assistance in obtaining specimens. We also thank Gary Hendrickson and Jerry Allen of Humboldt State University for their comments and guidance. LITERATURE CITED Ayres, W. O. 1854. Daily Placer and Transcript, May 30. Cited in: 1857 Proc. Calif. Acad. Sci. (1854-1857) 1:3-22. Burns, J. W. 1966. Western sucker. Pages 516-517 in A. Calhoun (Ed.), Inland fisheries management. Calif. Dept. Fish and Game. 546 pp. Eddy, S. 1969. How to know the freshwater fishes. Wm. C Brown Co., Dubuque, I A. 265 pp. Ferris, S. D., and C. S. Whitt. 1978. Phylogeny of tetraploid catostomid fishes based on loss of duplicate gene ex- pression. Syst. Zool. 27:189-206. Fowler, H. W. 1913. Notes on catostomid fishes. Proc. Acad. Natur. Sci. Phila. 65:45-71. Fuiman, L. A. 1978. Descriptions and comparisons of northeastern catostomid fish larvae. M. S. Thesis. Cornell Uni- versity, Ithaca, NY. 110 pp. Green, D. E. 1978. Analyzing multivariate data. The Dryden Press, Hinsdale, IL. 519 pp. Hubbs, C. L and C. Hubbs. 1953. An improved graphical analysis and comparison of series of samples. Syst. Zool. 2:49-57. Hubbs, C. L, and K. F. Lagler. 1958. Fishes of the great lakes region. Univ. Michigan Press, Ann Arbor, Ml. 213 p. COMPARISON OF SACRAMENTO SUCKER 1 87 Hubbs, C. L, and A. Perlmutter. 1942. Biometric comparison of several samples, with particular reference to racial investigations. Am. Natur. 76:582-592. Lee, D. S., C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAllister, and J. R. Stauffer. 1980. Atlas of North American freshwater fishes. North Carolina St. Mus. Natur. Hist. 854 p. Martin, M. 1 967. The distribution and morphology of the North American catostomid fishes of the Pit River system, California. M. A. Thesis. Sacramento State University, Sacramento, CA. 67 p. Mayr, E. 1969. Principles of systematic zoology. McGraw-Hill Book Company, New York, NY. 428 p. Moyle, P. B. 1976. Inland fishes of California. Univ. California Press, Berkeley, CA. 405 p. Ott, L. 1977. An introduction to statistical methods and data analysis. Duxbury Press, North Scituate, MA. 730 p. Robins, C. R., R. M. Bailey, C. E. Bond, J. R. Brooker, E. A. Lachner, R. N. Lea, and W. B. Scott. 1980. A list of common and scientific names of fishes from the United States and Canada, 4th Ed. American Fisheries Society Spec. Publ. No. 12. 174 p. Rutter, C. 1908. The fishes of the Sacramento-San Joaquin basin, with a study of their distribution and variation. Bull. U. S. Bur. Fish. 27:103-152. Smith, G. R. and R. H. Koehn. 1 971 . Phenetic and cladistic studies of biochemical and morphological characteristics of Catostomus. Syst. Zool. 20:282-297. Snyder, J. O. 1908. The fishes of the coastal streams of Oregon and northern California. Bull. U. S. Bur. Fish. 27:153-189. 1913. The fishes of the streams tributary to Monterey Bay, California. Bull. U. S. Bur. Fish. 32:49-72. 1914. The fishes of the streams tributary to Tomales Bay, California. Bull. U. S. Bur. Fish. 34:377-381. Wilson, E. O. and W. L. Brown, Jr. 1953. The subspecies concept. Syst. Zool. 2:97-111. Wishart, D. 1979. Clustan 1C users manual. California State Universities and Colleges, Los Angeles, CA. 136 p. 188 CALIFORNIA FISH AND GAME Calif. Fish and Came 73(3): 188-191 1987 EXTENT OF HUMAN-BEAR INTERACTIONS IN THE BACKCOUNTRY OF YOSEMITE NATIONAL PARK1 Bruce C. Hastings 2 and Barrie K. Gilbert Department of Fisheries and Wildlife Utah State University Logan, Utah 84322 A survey of human-bear interactions was conducted in 6 popular backcountry sites in Yosemite National Park during May-October 1979. Almost half of the 1647 parties interviewed noted bear activity (e.g. interactions, damages, or other evidence of bear presence) within the previous 24 hours, 41% directly interacted with bears, and over 12% sustained bear damage. Use of ranger patrols, advanced food storage techniques, and smaller party size appeared to reduce conflicts. Convenience for re- porting bear incidents was found to be highly relevant to management estimates of bear-related conflicts. In addition, expensive losses were much more likely to be re- ported to the National Park Service than minor damages. Research and management recommendations are discussed. INTRODUCTION Conflicts between bears and national park users have received considerable attention in recent years. Although serious problems with grizzly bears, Ursus arctos, are often more publicized, interactions involving black bears, Ursus americanus, are more common. Herrero (1985) determined that at least 500 people were injured by black bears in North America from 1960 to 1980. Prop- erty damage can also be substantial. Although improvements in management strategies (e.g. relocation of problem bears combined with intensive enforce- ment of regulations and better education of visitors) reduced problems in many frontcountry areas, increased recreational use of the backcountry during the 1960's and 1970's led to more frequent human-bear conflicts. Similar problems increased in the backcountry of Yosemite National Park during the mid-1 970's (Harms 1980). These problems led to an investigation of human-bear interac- tions and methods to separate the 2 species. Upon initiation of this project, it be- came apparent that no reliable information was available on the scale or extent of the problem in the backcountry. Thus, a portion of the study was directed to determining the proportion of backcountry parties directly affected by black bears. This paper analyzes the extent of the bear problem in several heavily used backcountry sites and makes relevant management recommendations. METHODS Informal interviews were conducted with overnight visitors at 6 popular backcountry sites during May through October 1979. These campgrounds were selected due to histories of human-bear interactions and applicability to other study objectives (see Hastings, Gilbert, and Turner 1981). During interviews, data collectors wore normal backpacking clothes. They ap- proached visitors mostly during mornings, introduced themselves as bear re- searchers and asked whether, within the previous 24 hours, anyone in the party had seen or heard a bear (interaction), sustained property damage from bears, 1 Accepted for publication April 1987. 2 Current address: Department of Forestry, Wildlife, and Fisheries, University of Tennessee, Knoxville, TN 37901-1071. HUMAN-BEAR INTERACTION 189 or noticed any bear activity (which included interactions, damages, or any other evidence that a bear had been in their personal or camp vicinity) . If they had re- ceived damage, they were asked to estimate the financial loss. Data were re- corded on small, unobtrusive notepads and transferred daily to data sheets. In- terviewers also recorded the size of each camping party; those with 8 or more persons per party were categorized as organized groups, usually consisting of 8 to 40 youths and a few adult leaders. In order to reduce bias when analyzing re- porting rates, visitors were informed at the conclusion of each interview that they had not officially reported any damage and that they should do so. All statistical comparisons were made with the Chi-Squared Test. RESULTS AND DISCUSSION A total of 1647 parties were interviewed (Table 1 ), which represented almost 90% of the 1841 groups camping at the study sites on days of interviews. One hundred and one of these parties consisted of 8 or more members. TABLE 1. Summary of Interviews for All Sites Studied, Yosemite National Park, 1979. No. of camps No. of camps % of camps with < 8 % of camps with > 8 % of camps Number of camps: Total interviewed visitors interviewed visitors interviewed At sites 1841 _ 1736 _ 105 _ Interviewed 1647 100.0 1546 100.0 101 100.0 With activity 793 48.1 725 46.9 68 67.3 With interactions 675 41.0 611 39.5 64 63.4 With damages 208 12.6 171 11.1 37 36.6 Damage estimates $2376 $1893 $483 Bears and people came into close proximity far more often than had been an- ticipated. For data combined from all 6 study sites, almost half (48.1%) of the parties noted bear activity within the previous 24 hours, 41 .0% experienced in- teractions with bears, and 12.6% sustained bear damage. Financial losses aver- aged $11.42 for each of the 208 parties experiencing damage. Although a single bear could occasionally be noticed by numerous campers at a congested site, most campgrounds were frequented by more than one bear, and each bear would often visit the site more than once a day. The reported levels of bear prob- lems were not the result of one bear per campground visiting the area only once per day. Almost two-thirds of the above data were collected in Little Yosemite Valley ( LYV) . LYV was previously known for its frequent and severe human-bear prob- lems, and received the highest human use in Yosemite's backcountry during 1 979 (NPS records). High overnight visitor use has been correlated with high prob- ability of bear conflicts in several parks (Merrill 1978, Singer and Bratton 1980, Keay and van Wagtendonk 1983). However, bear problems occurred less fre- quently in LYV than in other study areas (Tables 2-3). Other sites had signifi- cantly more parties (P < 0.001 ) reporting bear activity, interactions, or damages than LYV. In addition, the average financial loss for a bear damage incident was almost half as expensive in LYV. Reduced bear problems in LYV were probably due to the frequent ranger patrols and the use of somewhat advanced food stor- age techniques, such as metal bear-proof food containers. The major point is that a large percentage of backcountry users were inter- acting with bears and sustaining damages to food and equipment. Even LYV users were experiencing an unacceptable level of bear activity. The reader should be 1 90 CALIFORNIA FISH AND CAME aware, however, that data collection was biased toward areas known for bear ac- tivity. This was offset somewhat by the clumping of visitors in certain areas, and the corresponding concentration of problem bears wherever there were con- centrations of backcountry people ( i.e. at the same backcountry sites where peo- ple tended to concentrate). During 1979, 60% of the backcountry visitors in Yosemite used only four trailheads (van Wagtendonk 1981 ). Thus, estimates from this study are probably reasonable for numerous sites in Yosemite, espe- cially those with relatively heavy human use. However, sites with low human vis- itation may have been used less by problem bears. TABLE 2. Summary of Interviews for Little Yosemite Valley, Yosemite National Park, 1979. No. of camps No. of camps % of camps with < 8 % of camps with > 8 % of camps Number of camps: Total interviewed visitors interviewed visitors interviewed At sites 1159 _ 1102 _ 57 _ Interviewed 1040 100.0 986 100.0 54 100.0 With activity 405 38.9 377 38.2 28 51.9 With interactions 308 29.6 283 28.7 25 46.3 With damages 62 6.0 53 5.4 9 16.7 Damage estimates $436 $355 $81 TABLE 3. Summary of Interviews for All Sites Studied Except Little Yosemite Valley, Yosemite National Park, 1979. No. of camps No. of camps % of camps with < 8 % of camps with > 8 % of camps Number of camps: Total interviewed visitors interviewed visitors interviewed At sites 682 _ 634 _ 48 _ Interviewed 607 100.0 560 100.0 47 100.0 With activity 388 63.9 348 62.1 40 85.1 With interactions 367 60.5 328 58.6 39 83.0 With damages 146 24.1 118 21.1 28 59.6 Damage estimates $1940 $1538 $402 Large visitor groups were more likely to note bear activity, experience inter- actions, and especially receive damages (P < 0.001 ) (Tables 1-3). This is par- tially due to more people being available to interact with and lose food to bears. However, organized groups rarely appeared to be truly "organized." These groups were well-known for storing food improperly (e.g. tying foodsacks to tree trunks) and for violating NPS regulations concerning maximum group size. In order to better understand the extent of bear damages in the backcountry, data for LYV were again compared to sites without rangers. In LYV, 27.9% of the estimated number of parties experiencing damages reported those damages to the National Park Service, although 51 .6% of the estimated financial losses in dol- lars was reported (i.e. more expensive damages were reported more often). For backcountry areas without rangers on duty, only 1 .3% of the number of damages were reported while 2.8% of the financial losses were reported. Visitors ap- peared reluctant to report bear damage unless it was easy to do so; convenience for reporting bear conflicts has been described elsewhere as an important factor in improving reporting rates (Petko-Seus 1985). In addition, unwillingness to in- form the National Park Service of bear damages may have been generated by a fear of receiving a citation for improper food storage. However, these tendencies decreased with more expensive damage as seen by the much higher percent of estimated dollar loss being reported. People who had large losses appeared to be HUMAN-BEAR INTERACTION 191 more angry and more willing to risk a citation in hopes of being reimbursed by the Park Service for damages sustained. CONCLUSIONS AND RECOMMENDATIONS Large numbers of backcountry campers in 6 heavily used campgrounds in Yosemite National Park experienced interactions with black bears (41.0%) and sustained losses to food and equipment (12.6%). This does not mean that all in- teractions were unpleasant or inappropriate; seeing bears was obviously the most exciting feature of many visitors' trips. However, most visitors were not prepared for these situations. More parks should monitor the extent of bear problems and then relay those findings to visitors; if visitors recognized that one-fourth of the parties in certain areas receive bear damages, they might visit less heavily used areas or prepare better for dealing with bears. The legal size of organized groups in black bear habitat should be enforced. Perhaps the optimal leader : member ratio should be evaluated and regulated in order to make these parties truly "organized." Visitors using sites not patrolled regularly by backcountry rangers nor equipped with the most advanced food storage devices experienced more se- vere problems with bears. Obviously, the National Park Service cannot afford to place rangers and expensive equipment at many backcountry sites, but any in- crease in ranger/backpacker contact and improvements in bear-proofing tech- niques at sites known for bear problems could help. Convenience for reporting bear conflicts should be improved for backcountry users. Additional research on this subject could provide information on the nature of human-bear relationships to many backcountry managers. Sampling might be al- tered in the future to include large numbers of sites while keeping costs reason- able by interviewing campers at trailheads. Projects can be designed to evaluate the extent of problems throughout a wilderness area or to test effectiveness of food protection or aversive conditioning techniques at specific sites. In addition, managers can use information obtained with this interviewing approach to es- tablish indices for damages or other human-bear conflicts. ACKNOWLEDGMENTS This study was funded by the National Park Service, Utah State University Min- eral Lease Funds, and Utah State University Ecology Center. We thank all those who contributed their time and effort. We are especially grateful to M. Cherry, S. McGrew, J. Pearson, J. Picton, M. Poyadue, and R. Strong for collecting data, and S. Hastings for typing and editing this manuscript. LITERATURE CITED Harms, D.R. 1980. Black bear management in Yosemite. Int. Conf. Bear Res. and Manage. 4:205-212. Hastings, B.C., B.K. Gilbert, and D.L. Turner. 1981. Black bear behavior and human-bear relationships in Yosemite National Park. Nat. Park Serv. Tech. Rep. 2, Davis, Calif. 42 p. Herrero, S. 1985. Bear attacks: their causes and avoidance. Winchester Press, Piscataway, New Jersey. 287 p. Keay, J.A., and )..W. van Wagtendonk. 1983. Effect of Yosemite backcountry use levels on incidents with black bears. Int. Conf. Bear Res. and Manage. 5:307-311. Merrill, E.H. 1978. Bear depredations at backcountry campgrounds in Glacier National Park. Wildl. Soc. Bull. 6:123-127. Petko-Seus, P. A. 1 985. Knowledge and attitudes of campers toward black bears in Great Smoky Mountains National Park. Thesis, Univ. Tenn., Knoxville. 173 p. Singer, F.J., and S.P. Bratton. 1980. Black bear/human conflicts in the Great Smoky Mountains National Park. Int. Conf. Bear Res. and Manage. 4:137-140. van Wagtendonk, J.W. 1981. The effect of use limits on backcountry visitation trends in Yosemite National Park. Leisure Sci. 4:311-323. 192 CALIFORNIA FISH AND GAME Calif. Fish and Came 73 ( 3 ) : 1 92 1 987 NOTES FIRST OREGON RECORD FOR THE COWCOD, SEBASTES LEVIS A cowcod, Sebastes levis, was recorded for the first time along the coast of Or- egon on 19 April 1986, approximately 65 km west of Newport (lat44°36.TN, long 124°38.7'W). The specimen, an immature male, measuring 227 mm standard length and weighing 318 g, was captured in 256 m by the JUL-E (skippered by Wade E. Richardson) while trawling for rockfish with a 9.5 cm mesh otter trawl with roller gear. The specimen is deposited in the fish collection of the Depart- ment of Fisheries and Wildlife, Oregon State University, OS 11063. This collec- tion extends the northern range of Sebastes levis into Oregon waters some 534 km from Usal, California, which was recorded as the northern limit by Odemar (1964). We would like to extend our appreciation to C. Harding, R. Demory, and W. Barss for reviewing this manuscript. We also wish to thank the skipper and crew of the JUL-E. This work was supported by Oregon State University Sea Grant Pro- gram (Grant No. NA85AA-D-SG095, Project No. R/ES-7), National Marine Fish- eries Service (Contract NA-85-ABH-00025), and Oregon Department of Fish and Wildlife. Technical paper 7936 of the Oregon Agricultural Experimental Sta- tion. LITERATURE CITED Odemar, M. W. 1964. Northern range extension of the cow rockiish, Sebastodes levis. California Fish and Came, 50(4) :305. — Daniel L. Erickson and Ellen K. Pi kite h, Mark O. Hatfield Marine Science Cen- ter, Department of Fisheries and Wildlife, Oregon State University, Newport, Oregon 97365. Accepted for publication January 1987. Pholoelectronic composition by CALIFORNIA OFFICE OF STATE PRINTING INSTRUCTIONS TO AUTHORS EDITORIAL POLICY California Fish and Game is a technical, professional, and educational journal devoted to the conservation and understanding of fish and wildlife. Original manu- scripts submitted for consideration should deal with the California flora and fauna or provide information of direct interest and benefit to California researchers. Authors may submit an original plus two copies, each, of manuscript, tables, and figures at any time. MANUSCRIPTS: Authors should refer to the CBE Style Manual (Fifth Edition) and a recent issue of California Fish and Came for general guidance in preparing their manuscripts. Some major points are given below. 1. Typing — All material submitted, including headings, footnotes, and literature cited must be typewritten doublespaced, on white paper. Papers shorter than 10 typewritten pages, including tables, should follow the format for notes. 2. Citations — All citations should follow the name-and-year system. The "library style" will be followed in listing literature cited. 3. Abstracts — Every article must be introduced by a concise abstract. Indent the abstract at each margin to identify it. 4. Abbreviations and numerals — Use approved abbreviations as listed in the CBE Style Manual. In all other cases spell out the entire word. TABLES: Each table should be typewritten with the heading margin left justified. Tables should be numbered consecutively beginning with "1" and placed together in the manuscript following the Literature Cited section. Do not double space tables. See a recent issue of California Fish and Game for format. FIGURES: Consider proportions of figures in relation to the page size of California Fish and Game. The usable printed page is 1 1 7 by 1 91 mm. This must be considered in planning a full page figure, for the figure with its caption cannot exceed these limits. Photographs should be submitted on glossy paper with strong contrasts. All figures should be identified with the author's name in the upper left corner and the figure numbers in the upper right corner. Markings on figures should be made with a blue china marking pencil. Figure captions must be typed on a separate sheet headed by the title of the paper and the author's name. PROOFS: Galley proofs will be sent to authors approximately 60 days before publi- cation. The author has the ultimate responsibility for the content of the paper and is expected to check the galley proof carefully. PAGE CHARGES AND REPRINTS: All authors will be charged $35 per page for publication and will be billed before publication of manuscripts. Reprints may be ordered through the editor at the time the proof is submitted. Authors will receive a reprint charge schedule along with the galley proof. Jo > ^ -n © a c» m -i » i 5 Co c CO w ■ol£ Zi p o K RA POST, AID "o n > -» 4*. o o m «o — m