Recruitment, Dispersal, and Demographic Rescue in Spatially-Structured White-Tailed Ptarmigan Populations ’

The Condor 102:503-516 D The Cooper Ornithological Society Zoo0

RECRUITMENT, DISPERSAL, AND DEMOGRAPHIC RESCUE IN SPATIALLY-STRUCTURED WHITE-TAILED PTARMIGAN POPULATIONS ’

KATHY MARTIN Canadian Wildlife Service, 5421 Robertson Rd, RR I Delta, British Columbia, Canada, V4K 3N2, and Department of Forest Sciences, 2424 Main Mall, University of British Columbia, Vancouver, British Columbia, Canada, V6T 124, e-mail: [email protected]

PETER B. STACEY Department of Biology, University of New Mexico, Albuquerque, NM 87131, e-mail: [email protected]

CLAIT E. BRAUN* Colorado Division of Wildlife, 317 West Prospect Rd., Fort Collins, CO 80526, e-mail: [email protected]

Abstract. We studiedrecruitment and dispersalof White-tailed Ptarmigan (Lagopus leucurus) breeding in naturally fragmented alpine habitatsat four study sites in Colorado from 1987-1998. Almost all recruitment for both sexes, particularly females, was of birds produced outside local populationsand also external to nearby studiedpopulations. Populations were more dependent on female recruitment than on male recruitment to sustainthem, and patterns of recruitment were not correlated with local survival of adults or production of young the previous year, except at one site for females. Over 95% of recruitswere yearlings. Breeding dispersal of adults, an infrequent but regular event, was also important to inter- populationconnectivity. Our data for multiple populationsallowed us to describemovement patterns among populations to assess consistency with conditions required for a rescue system. After widespread reproductive failure in one year, we expected all populationsthe next year would have low recruitment due to a reduced supply of recruits produced in the region. Recruitment was low, but impact varied among populations.We conductedan over- winter study of radio-marked offspring to determine possible influences of winter site location and relatives on recruitment patterns.Contrary to expectation,offspring remained on or near breeding sites in winter, but were not located near their mothers or siblings. Re- cruitment location was related to winter site location. White-tailed Ptarmigan exhibit a well developed capacity for external recruitment that allows them to persist in small populations with stochasticconditions for breeding and survival. Extensive external recruitment may be a general pattern for birds. Kev words: demographic rescue, dispersal, Lagopus leucurus, local survival, metapo- pulazon processes, recruitment, White-tailed Ptarmigan.

INTRODUCTION and Kodric-Brown 1977). Stacey and Taper Periodic rescue of declining populations through (1992) modeled the importance of immigration recruitment from productive populations (dis- rescue to persistence of small populations, and persal rescue) may be an important feature of showed that external recruitment was vital to en- suring population persistence and stability. Dis- avian population biology. Dispersal rescue de- persal rescue is predicted to work most effec- scribes the processwhereby populations that ex- tively when neighboring populations are in close perience one or several poor years of reproduc- proximity, but are uncorrelated in patterns of an- tion or survival escapeextinction or a population nual production and survival (Stacey et al. bottleneck by immigration of individuals from 1997). If inter-patch dispersal is high, local ex- elsewhere in the metapopulation system (Brown tinction may not occur (Harrison 1991). Thus, dispersal may operate regularly to ensure local I Received 23 April 1999. Accepted 3 1 March 2000. population stability; extinction may be a failed 2 Current address:5572 No. Ventana Vista Rd., Tuc- rescue process. son, AZ 85750. Studies of the patterns and mechanisms of re-

[5031 cruitment and dispersal in multiple populations lated across sites and years, and that small pop- allow a direct assessmentof the importance of ulations remained stable despite stochasticity in demographic connectivity among populations. internal population dynamics (Martin et al. Metapopulation processesmay be influenced by 1997, K. Martin, unpubl. data). Here, we present both local and/or regional stochasticity (Hanski data on origins and patterns of recruitment for 1991, 1994, Hanski and Gilpin 1991, Harrison four study populations that showed strong reli- 1991). A knowledge of synchrony in popula- ance on external rescue. For recruits, we present tions within a region is crucial to understanding data on the relative amounts of natal (juveniles) population and multi-population persistence versus breeding dispersal (breeding birds be- times (Harrison and Quinn 1989, Hansson tween years). Settlement patterns in relation to 1991). natal territory are given for offspring recruits in Regardless of whether a metapopulation one population. structure can be demonstrated for a species, in- We also present data on recruitment and pop- vestigating metapopulation processes such as ulation size in a year following widespread re- dispersal and rescue, and how they interact with productive failure, and compare these to recruit- stochastic factors, is vital to understanding the ment patterns from all previous years when re- population dynamics for species, particularly in productive failure varied spatially across popu- fragmented habitats. Alpine habitats, given their lations. Thus, with this natural experiment, we naturally patchy character with regular stochas- examine patterns of recruitment caused by local tic factors, are ideal for examining rescue pro- and regional events. We anticipated that all pop- cesses. Dispersal and other population transfer ulations would have low recruitment in the year mechanisms for vagile vertebratesare poorly un- following regional-level failure due to a reduced derstoodand logistically difficult to study at rel- supply of recruits in the region, but that popu- evant biological and spatial scales (Martin lations with generally higher local survival of 1998). Data are usually lacking on social inter- adults would be impacted less than populations actions or movements in the post-breeding sea- normally requiring a high rate of rescue. We ex- son, factors considered to be important to off- pected that one population that normally recruit- spring making “decisions” regarding dispersal ed a higher proportion of adult females would and recruitment. Most life history studies as- be buffered from this regional-level event. Fi- sume that patterns of offspring settlement show nally, with a telemetry study, we examined pos- a gradual decline away from natal territories, but sible social and behavioral mechanisms for natal there are logistical and analytical problems in recruitment. In one year, we followed radio- estimating numbers and distance of widely dis- marked offspring over winter to the following persing offspring (Koenig et al. 1996, Nichols spring, and examined influences of movements 1996). by mothers and siblings and winter dispersal be- To examine the importance of metapopulation havior on offspring recruitment patterns. processesin naturally fragmented habitats, we studied four populations of White-tailed Ptar- METHODS migan (Lugopus Zeucurus) in Colorado. About STUDY AREAS 95% of both sexes spent their entire breeding life in one population, satisfying the definition We studied four populations of White-tailed of a discrete population suggestedby Harrison Ptarmigan breeding in naturally fragmented high (1991), and patcheswere connected by natal dis- alpine habitats in the Colorado Rocky Moun- persal. Our populations showed dramatic varia- tains (39”34-40’N, 105”35-53’W) from 1986 to tion in offspring production and return (local 1998 (Fig. 1). The life history and breeding bi- survival) of adults acrossyears and sites (Braun ology of White-tailed Ptarmigan are detailed in et al. 1993, Hannon and Martin 1996). Repro- Braun and Rogers (1971) and Braun et al. (1976, ductive failure appeared driven by climatic 1993). Each of the four study areas comprised events and internal ecological processessuch as individual mountain patches separatedfrom the depredation of eggs and young (Braun et al. other areas by distances of l-20 km, and in 1993, K. Martin and K. L. Wiebe, unpubl. data). some directions by areas of unsuitable ptarmigan Preliminary results indicated that variation in re- breeding habitat (forest, wetland, willow Mix productive success and survival were uncorre- spp. shrub). The Mt. Evans (ME) study area was Mt

Otter MOUtltill Mt Wilcox 8 -20 km to NW GriZZ%’y Gray Wolf @a MOUtltitl

Rosalie Peak I 2km

Mountain FIGURE 1. Study sites for White-tailed Ptarmigan in Colorado, including Guanella Pass winter site and adjacent mountainsthat were surveyed less intensively in late summer.ME-Mt. Evans, GP-Guanella Pass,ST- Square Tops, and LP-Loveland Pass. The shaded lines on ME study area indicate the four major drainages used to evaluate philopatry of internal recruits.

13.2 km2 in size, and from 1987 to 1996 annual were missing for some years including 1995, the spring population sizes ranged from 26 to 67 year of regional failure. In 1997 and 1998, we birds (Fig. 1, Table 1). Guanella Pass (GP), 6-7 conducted less intensive censusing on ME and km from ME, comprised about 2.8 km*, and ST, and present data only for our over-winter from 1987 to 1996 (excluding 1988), popula- study of radio-marked birds. tions varied from 6 to 20 birds. Square Tops (ST) was l-2 km from GP and 8-11 km from FIELD PROCEDURES ME. Square Tops was 4.75 km2 and during 1992 Field methods and data collection protocols to 1996 spring populations varied from 13 to 21 were identical for all study sites from 1987 to birds. Loveland Pass (LP), about 2.7 km2, was 1996, and are described in Artiss and Martin about 20 km west of ST Population sizes ranged (1995). In spring, pairs and single birds were from 12 to 31 birds in 1987, 1989-1992, and located using playbacks of male territorial calls 1994. We also visited 12 mountains adjacent to or with dogs, and captured using noose poles, our study sites in late summer, and captured and noose carpets, or nets. We did not capture fe- marked the ptarmigan adults and chicks we en- males until they had spent several days with a countered. Although sampling of populations on particular male on a territory and the pair bond adjacent mountains was limited, it allowed an appeared stable. We affixed necklace design ra- opportunity to measure dispersal and recruit- dio transmitters (3.4 g, Holohil Ltd., Carp, On- ment between our studied populations and ad- tario, Canada) on all females and some males to jacent sites in the region. The number of years facilitate relocation and to monitor mating status of data reported vary for different aspectsof the and breeding success.Birds were classified as study. We present less detail for LP becausedata yearlings (< 1 year old) or adults (> 1 year) based on the amount of pigmentation on pri- of dispersal for offspring produced in our pop- maries 8 and 9 (Braun and Rogers 1971). All ulations. In 1997, we radio-marked chicks and study populations contained a high proportion of brood hens in late summer. Eighteen of 21 radios known-age marked birds, with the exception of had 1 year battery life (wt = 5.8 g; about 1.5% the first study years for GP and LP (1987) and of ptarmigan post-breeding body mass; three had ST (1992). For about 95% of 196 females mon- mortality sensors), and radios on three brood itored, we obtained data on nesting successand hens had 4 month battery life. We recorded sig- number of fledglings (young 2 25 days old) pro- nal locations for radio-marked birds using fixed- duced in a season. For all 369 chicks reported wing aircraft during four winter flights (24 No- here, natal population was determined because vember 1997, and 6 January, 16 March, and 28 chicks were individually banded or wing-tagged April 1998) and one summer flight (11 June before brood breakup in late summer; in most 1998). We flew over ME, ST, GP and LP several cases natal territory could be assigned. From times and then censusedall suitable alpine hab- 1989 to 1993, about 74% of 117 chicks were itat around the study areas extending up to a marked with small numbered aluminum patagial distance of 100 km to locate dispersedbirds. We tags within a few days of hatch. Loss of patagial recorded locations, directions, and distances of tags was low for ptarmigan (Hannon et al. 1990, offspring from their natal territories. In summer K. Martin, unpubl. data). 1998, we did ground checks of radio-marked Over the study, all birds located were cap- offspring to ascertain survival, sex, and breeding tured and individually color-marked. We relo- status. cated White-tailed Ptarmigan individuals throughout the breeding seasonto determine sur- DATA ANALYSIS vival, mating status, and reproductive success. Mean values are presentedwith ? SE. Statistical Return of birds that bred in a previous year to significance is reported for two-tailed tests (P < their original study population, an estimate of 0.05) using ANOVA, and log likelihood chi- local survival, was determined by thorough square tests, with G-tests when sample sizes checks of all the study areas and adjacent sites were small. Spearman rank correlations (rs) were throughout the summer to encounter banded done to examine synchrony of recruitment pat- birds that had dispersed. Preliminary mark-re- terns with chick production in the previous year capture analyses indicated that we located about and patterns of local survival. Given unequal 95% of birds resident on study areas, and thus populations sizes and different years of study for our data on “returns” closely approximate local sites, we do not report global correlations for all survival (K. Martin, unpubl. data). Banding re- populations. cords and age classesof birds were used to clas- sify birds as returns or recruits. Birds were clas- RESULTS sified as “returns” if they were banded and PATTERNS OF RECRUITMENT WITHIN AND known to have resided on that study area in a AMONG POPULATIONS previous year. A small number of adults changed Over all study sites, we marked 369 ptarmigan sites between years (i.e., breeding dispersal), and chicks (156 chicks leg-banded and 213 wing- if birds dispersedgreat distances from the study tagged) from which we could measure patterns areas they were unlikely to be detected (Wiebe of recruitment and dispersal from 1987 to 1996. and Martin 1997). Birds were classified as “re- From 1987 to 1996, we monitored local survival cruits” when first observed on a study area, ei- and reproductive successfor 292 female-years ther as marked or unmarked yearlings (i.e., first (196 individual females). New recruits account- time breeder) or as individually-marked adults ed for about 50% of resident breeding females that were observed previously on other areas. over the study, with proportions ranging from 0 Recruits were classified as internal (produced in to 100% acrossyears and study populations (Ta- that study population), local (produced in anoth- ble 1). ME had the lowest average proportion of er study population or from an adjacent moun- recruits (45.2%) among populations, and GP the tain), or external (produced outside our studied highest with an average of 69.2% of females in populations). the population consisting of recruits. ME, the We conducted a telemetry study of juvenile largest population, had the lowest annual varia- movements to assesspatterns and spatial scale tion in proportion of recruits (30-63.6%) among

TABLE 2. Recruitment and age of recruits in three study populations of White-tailed Ptarmigan in Colorado in 1996 after harsh environmental conditions in 1995, versus all previous years.

1996 Previous years Population Recruits Recruitsa size Population size n n % adult mean + SE Mean + SE n % adult

Females Mt. Evans 1.5 5 20 22.1 5 1.7 10.2 ir 1.1 93 2.2 Guanella Pass 2 2 0 6.2 ? 0.6 4.2 -t 1.0 2.5 4.0 Square Tops 5 3 67 6.3 ? 0.5 3.3 2 0.7 10 50.0 Males Mt. Evans 20 7 57 22.2 i 2.2 8.0 + 0.7 72 2.8 Guanella Pass 4 0 9.2 k 0.4 2.5 t 0.6 15 13.3 Square Tops 10 3 67 10.3 t 1.8 1.0 + 0.6 3 0.0

a Recruitment from previous years includes: ME-1987 to 1995; GP-1990 to 1995; ST-1993 to 1995 breeding females. Smaller populations had high- tion as 50% of females recruiting there were ob- er annual variation in their dependence on re- served first as adults (> 1 year old). cruits, with GP showing extremes of 0 to 100% Despite working on four populations, three in recruits. Numbers of female recruits in a given close proximity, only 29 marked offspring (8 fe- year were not correlated with number of return- males and 21 males) recruited to our study sites ing females (local survival) for ME (rs = 0.24, between 1987 and 1996 (Fig. 2). Numbers of P = 0.51, rz = 10 years) or GP (rs = -0.28, P known-origin recruits varied from none to six = 0.54, n = 7 years). Numbers of female re- annually (Table 1). We were able to determine cruits in study populations were correlated with the origin of only 8 (5.9%) of 137 female ptar- chick production in the previous year at GP (rs migan recruits to our populations from 1987 to = 0.94, P = 0.001, n = 7 years), but not at ME 1996 (Table 1, Fig. 2). Seven of eight female (I, = 0.59, P = 0.07, n = 10 years). recruits of known origin were yearlings. Only 5 From 1987 to 1996, we monitored local sur- of 98 females recruiting to Mt. Evans were pro- vival and reproductive success for 347 male- duced on that site, and none was known to have years (168 males, includes paired and unpaired originated from the other study populations or males). Recruits comprised only 30.6% of male adjacent mountains (Table 1). Two of 27 recruits populations, a significantly lower proportion to GP were produced on other study sites (Fig. than females (G, = 25.1, P < 0.001). Propor- 2, ME, ST). We observed local recruitment of tions of male recruits varied across study pop- females (connectivity) between all three proxi- ulations from 14% at ST to 35.8% at ME (Table mate populations (ME, GP, and ST), but none 1). Thus, males showed the opposite pattern of with Loveland Pass (Fig. 1). Likewise for males, dependence on recruitment to females, with the the majority of recruits came from outside our larger population (ME) showing the highest proportion of recruitment. As for females, the num- studied populations. However, male natal phil- ber of male recruits in a given year was not cor- opatry was higher than females as 21.4% of related with numbers of returning males (local male recruits were internal recruits (Fig. 2; survival) at ME (T$ = -0.4, P = 0.69, n = 10 males vs. females: G, = 15.0, P < 0.001). At years) or at GP (Y, = -0.05, P = 0.92, n = 7 ME, 25% of 77 male recruits were produced years). Also, numbers of male recruits to study there, and there was 1 local recruit from ST (Fig. populations were not correlated with the previ- 2). With our less intensive population data, we ous years’ chick production at ME (I; = 0.12, P encountered none of our marked offspring = 0.73, n = 10 years), nor at GP (rs = 0.28, P among the 26 birds captured and banded (14 = 0.55, IZ = 7 years). males, 12 females) on adjacent mountains, and In general, 95% of ptarmigan of both sexes none of 18 chicks tagged or banded on adjacent newly recruiting to populations were first time mountains recruited to our study sites. Also, we breeders (yearlings, Table 2). ST was an excep- recorded no marked offspring among 33 ptar- Females observed late spring dispersal where females 100 1 93 that settled on our study sites for one to several 80 weeks, desertedtheir mate and territory, and dispersed to other mountains. In 1990, this phe- 60 nomenon occurred at a higher rate and on all 40 study sites. In 1990, females paired normally, and then between 17 May and 6 June, 17% of 20 42 radio-marked females dispersed to sites lo- , Ol;;, 0 30+ km. The mean date of clutch initiation for 1990 was 10 June -C 0.7 for 31 females remain- % ing on study sites. Nine (64.3%) of 14 dispersing 2 Males females (populations and years pooled) were yearlings compared to the study average of % 100 46.9% yearling breeding females (1987 to 1996, g 80 n = 307 female-years, G, = 1.6, P = 0.2). In 60 two cases,females desertedmates and territories : three weeks and four weeks after settling, and moved 3-4 km to breed on adjacent mountains. They made these movements 5 days and l-2 days, respectively, before initiating egg laying. All females attempted breeding in their new lo- ME GP ST cation and their overall reproductive successof 2.6 chicks female-l (n = 8 females) was above FIGURE 2. Origin of birds recruiting to three White- tailed Ptarmigan populations in Colorado. Study site average for our studied populations (1.04 + 0.1 abbreviations as in Figure 1. Years of study varied with chicks female-‘, IZ = 159 female-years). Dis- site; ME: 1987-1996, GP: 1990-1996, ST 1993- persal of females, even when just before laying, 1996. Internal represents a recruit produced in that also did not appear to decreasesurvival. One of population, local recruits were produced in another nine spring-dispersingfemales located at the end study population, and external recruits came from outside the studied populations. Numbers above bars in- of the season was found dead, comparable with dicate sample size for recruits. the 12% mortality rate during the breeding season for females in our study populations (Braun et al. 1993). migan banded as either spring or fall migrants The proportion of unbanded adults recruiting on our study sites. to our populations was a measure of breeding Settlement patterns of internal recruits were dispersal between years because all females at- expected to be highest close to natal territories tempt breeding in their yearling year, and most with a gradual decline as one moves away. ME birds previously resident in our populations were was naturally divided from north to south into banded. About 6.2% of 128 females recruiting four major drainages (Fig. 1); 15 of 18 male to our populations were adults and thus repre- internal recruits settled in the same drainage as sented breeding dispersal (Table 2). For the their natal territory, and the remaining 3 settled above data, we excluded birds on edges of study in an adjacent region. In contrast, one of three sites where we might have missed recruits from female internal recruits at ME settled close to previous years. We have direct observations of her natal territory and the other two settled in breeding dispersal for four females (2.6%) out the most distant section from their natal site. of 155 between-year observations on our sites. Thus, male offspring returned close to their natal Two females moved from GP to elsewhere, one territory or they did not return, and female off- after two years of breeding there. Only one male spring settled away from natal territories. moved between study sites, recruiting to ST as a yearling and moving one year later to ME BREEDING DISPERSAL where he remained for four years. Thus, breed- We observed movements between populations ing dispersal was infrequent, but occurred reg- after natal dispersal events, both within season ularly, and was more common among females and between years. In four of seven years, we than males. PATTERNS OF RECRUITMENT AFTER A (mostly yearlings) to maintain its population of REGIONAL FAILURE IN REPRODUCTION six to eight pairs (Table 1). In 1996, GP showed Unseasonally late and protracted snowstorms the most dramatic population decline from 1995 through May, June, and early July 1995 resulted as both local survival and recruitment were the in the latest breeding season in our 1Zyear lowest observed in the study for both sexes (Ta- study. Weather records from Mt. Evans showed ble 1). ST was the only population to maintain that snowmelt in 1995 was the latest since ptar- its density in 1996; it did so by maintaining its migan studies began in 1966 (C. E. Braun, un- normal rate of return and recruitment for both publ. data). In 1995, we observed high repro- females and males (Tables 1 and 2). ductive failure in all populations with only 16% of 32 females raising chicks in 1995 compared TELEMETRY STUDY OF OFFSPRING to about 50% of 164 females in previous years RECRUITMENT: SOCIAL FACTORS AND OVER- WINTER MOVEMENTS (G, = 14.2, P < 0.001). Mean chick production of 0.4 -C 0.2 chicks female-’ (n = 32 females) To examine factors affecting natal recruitment, in 1995 was the lowest observed during the we conducted a telemetry study of juvenile ptar- study, and less than half that of previous years migan. In late August and September 1997, we (mean = 1.0 +- 0.1 chicks female-l, F,,,,, = 3.6, radio-marked 6 brood females and 12 chicks on P = 0.05, K. Martin and K. L. Wiebe, unpubl. ME and 1 brood female and 2 chicks at ST (1 data). In 1995, only one of seven radio-marked chick of each sex per brood where possible). On females that left the study areas in spring and winter census flights, we obtained 52 signal lo- bred on adjacent mountains successfully raised cations for 19 of 21 radio-marked birds; all chicks, suggestingthat breeding successin 1995 chicks were located (n = 40 locations for 14 also was low off the study areas. However, over- birds) and 5 of 7 brood hens (Table 3). The two winter survival of adults was not affected by the undetected brood females had 4-month radio- harsh 1995 season. Local survival was not re- tags. The most unexpected result was that most duced after the 1995 breeding season for either birds remained near, often within 1 km of breed- sex; 41% of 32 females returned to the study ing sites during winter. They did not remain on sites in 1996 compared to a return rate of 43% their natal territories but rather moved around of 265 female-years in previous years, and the area or to adjacent mountains. One ME chick 53.5% of 43 males returned in 1996 compared moved southwest 18 km to a winter site, and a to 65% of 204 male-years in earlier years (1996 second chick was located three times during vs. previous years: females: G, = 0.09; males: winter on or near ME, but in late April moved G, = 2.0, P = 0.15). 20 km to the northwest. In only 1 of 52 locations We used the harsh year of 1995 as an oppor- was a bird from ME found on the traditional tunity to assessthe possibility of a regional dis- Guanella Pass winter site. Offspring did not ap- persal rescue system, and expected that overall pear to be associating with parents or siblings recruitment rates should be much reduced in over winter (Table 3). 1996 following the widespread reproductive We recorded recruitment locations of radio- failure in 1995. As expected, female population marked offspring by an aerial censuson 11 June, sizes declined at two sites in 1996 (29% at ME followed by several ground checksbetween May and 75% at GP) from 1995 levels and for males and August 1998. Six chicks (five males) re- by 70% at GP (Table 1). The number of female cruited to their natal populations, one female recruits was lower than in any other year at ME from ME recruited to GP, and five chicks dis- and lower than in all other years but one at GP persed 9.5-20 km from their natal territories (Table 1). Male recruitment in 1996 did not (Table 3). We verified survival, sex, and breed- show the sharpdeclines observed in females, but ing status for all internal and local recruits, but over half the recruits in 1996 were adults in con- were less successful in obtaining this informa- trast to under 3% in other years (Tables 1 and tion for external recruits (likely predominantly 2; adult male recruits: 96 vs. previous years: G, females) as late spring snows hampered trips to = 25.0, P < 0.001). In 1996, 20% of female distant mountains in June 1998. For some dis- recruits were adults compared to an average of persed offspring, we determined locations from 6.3% in other years (Table 2, G, = 0.95). Over the air and assigned probable survival based on the study, GP depended heavily on recruits changes in signal locations during aerial checks TABLE 3. Winter migration, dispersal, and recruitment locations and behavior during 1997-1998 for radio- marked White-tailed Ptarmigan offspring in Colorado.

No. locations Recruitment No. on study Close to Close to Recruitment site close to Sex n locations area sibling? mother? sitea winter siteb

Male 5 18 18 o/12 1115 5 internal 515 Female 2 5 3 o/5 1 internal 212 1 local Unknownc 7 17 3 2116 0115 5 externala 3/5h Total 14 40 24 2133 l/30 6 internal 10l12b 1 local 5 externala

aExternal recruits dispersed 9.5-20 km to the south, southeast, southwest Cn = 2 chicks), and north of their natal areas. b Recruitment site data were not scored for two offsprmg because it is possible they were dead. c Sex was unknown for offspring that were not located by ground checks and line of sight checks from different ground ment, particularly for females, came from out- locations in June and August 1998. Despite lim- side the study populations and also from outside ited censusing in 1998, 30% of 30 chicks banded nearby populations. This pattern occurred in all in 1997 (includes radio-marked chicks) recruited years and applied to both sexes; external recruit- to our study populations (1 local and 8 internal ment varied annually from 78 to 100% for fe- recruits) in 1998, the highest recruitment rate re- males and from 37 to 100% for males. Another corded in our study (annual range: 0 to 18.2%), population study (1966-1989) of ptarmigan in and well above the overall recruitment rate of Rocky Mountain National Park reported 77% 7.9% for previous years (369 marked chicks, G, external recruitment for females and 50% for = 13.3, P < 0.001). males (Giesen and Braun 1993). It is possible We examined whether recruitment sites might that we underestimated internal recruitment as be related to winter dispersalmovements. For all we could have missed some locally produced internal and local recruits and three of the five offspring in some years. However, we recorded external recruits, recruitment sites were close to the reproductive fate of over 95% of resident where they spent the winter (Table 3). Radio- females in our populations, and we marked over marked offspring that dispersed had moved off 90% of chicks with brood females. We surely the natal area by late November, and recruited underestimated local recruitment because not all close to or on the site where they were located birds on adjacent mountains were marked. How- during winter. However, for one dispersing off- ever, the extremely low rate of recruitment of spring we recorded no movements over winter, those that were marked between the study pop- but in April, it dispersed 20 km north and then ulations is undeniable evidence that local re- settled 15 km to the south. Our data are minimal cruitment is infrequent, and that external recruit- estimates of movement because it is likely that ment dominates. most offspring made significant movements and Our ptarmigan breeding populations depended visited multiple areas before choosing breeding heavily on annual recruitment to sustain them, sites. especially for females where about half of the population each year was comprised of new DISCUSSION birds. Although recruitment varied dramatically EXTERNAL RECRUITMENT AND DISPERSAL over years, it was not strongly or consistently RESCUE correlated with internal population dynamics. White-tailed Ptarmigan are suitable to study me- Patterns of recruitment were not correlated with tapopulation processessuch as dispersal rescue local survival of either sex at the two sites with and recruitment. Both sexes are territorial and seven or more years of data. If population size they inhabit patchy environments with small was strongly influenced by local production, one populations and exhibit high demographic and would expect to see the strongest relationship environmental stochasticity. The most remark- between male recruitment and chick production able feature of our study was that most recruit- in the previous year, given higher natal philopatry of males. However, the only significant Hannon and Martin 1996). Natal philopatry was correlation between recruitment and chick pro- lower in British Columbia than in the Arctic duction in the previous year was observed for (Martin and Hannon 1987). Studying multiple females at GP and not for males. connected populations provides more accurate Annual variation in demographic factors, es- information on general patterns and variance in pecially if mediated by environmental stochas- recruitment that increase our understanding of ticity, has been suggestedto select for increased spatial population structure. For example, if we natal and breeding dispersal distances (Stenzel had studied only the Mt. Evans population, we et al. 1994, Paradis et al. 1999). Our ptarmigan would have missed about 5% of marked male populations showed surprising stability in pop- recruits and 37.5% of female recruits. Within the ulation size over our study. Despite small pop- same geographic region, recruitment rates have ulation sizes, low densities, relatively low fecun- been shown to vary widely within speciesin re- dity, and high annual variation in most popula- lation to patch size and proximity to other breed- tion parameters, none of our studied White- ing populations. The proportion of external re- tailed Ptarmigan populations have been cruits entering Song Sparrow (Melospiza melo- extirpated, and probably are not threatened. dia) populations varied from 5% on an island Most ptarmigan populations have persisted for that could sustain over 70 pairs, 57% on small at least 30 years (Braun et al. 1993), and records islands close to a mainland source, to 92% on of ptarmigan breeding on Mt. Evans extend back the mainland (Smith et al. 1996). Song Sparrow to 1890 (Lewis 1904). populations with higher rates of external recruit- We observed the majority of our recruits of ment were more stable, consistent with the con- known origin (predominance of males) to settle clusions of Paradis et al. (1999) that local syn- in their natal population. However, we did not chrony increased positively with dispersal dis- observe male offspring from ME or ST recruit- tances for 53 bird species in Britain. Although ing to GP, ST, or to adjacent mountains. Thus, investigators tend to focus more on philopatric male offspring returned close to their natal ter- offspring, these usually constitute less than 50% ritory or they did not return, suggesting a dis- of recruits (Greenwood 1980). Thus, extensive continuous pattern of offspring dispersal as op- external recruitment is probably a general pat- posed to a steady decay with distance from natal tern for birds. territory. Although we observed 75 to 100% external recruitment for our populations in the ear- PROPENSITY FOR DISPERSAL AND COSTS OF lier years of our study, we also have the puzzling DISPERSING result from our telemetry study that all radio- Movement between different breeding sites has marked male offspring we located returned to been documented within and between seasons their natal population in 1998. Although we fol- for many bird and mammal species(Greenwood lowed multiple populations in reasonable prox- 1980). Snowy Plovers (Charadrius alexandri- imity, our study, like most others, was not de- nus) in California have dispersed up to 175 km signed specifically to measure demographic ex- to breed between seasons; these movements change among sites. were not related to prior reproductive success The phenomenon of external recruitment has (Stenzel et al. 1994). About 5% of females been observed in other studies. Two single-site moved to another population after having com- studies of Willow Ptarmigan (Lagopus lugopus) pleted a breeding season elsewhere (Wiebe and showed different rates of external recruitment, Martin 1997). Breeding dispersal may be un- one in the central Canadian Arctic had about common but contributes significantly to inter- 5 1% external recruitment for males and 9 1% for population connectivity because it happens reg- females, whereas a population in northwestern ularly in all ptarmigan populations (Hannon and British Columbia had 89% external recruitment Martin 1996, this study). The high proportion of for males and 97% for females (Martin and Han- adult female recruits at Square Tops likely rep- non 1987). The two studies differed in spatial resents a different pattern of recruitment at that scale and population density with the arctic site as we feel our sampling efficiency at ST was study site being over twice the size of the site similar to other populations. Also, although un- in British Columbia (10 vs. 4 km2) but less than common, we observed late spring breeding dis- half the density (5-10 vs. 20-40 pairs ktn2, persal in most years. Interestingly, the phenomenon of late spring dispersal in 1990 was ob- from the study areas, and the potential areas for served also in Rocky Mountain National Park, dispersal were vast and often inaccessible. Giv- 150 km to the northwest, where after settling for en the spatial scales involved for most verte- one to several weeks, 3 of 12 radio-marked fe- brates, radio telemetry is required for survival male White-tailed Ptarmigan dispersed to other estimation in dispersal studies (Nichols 1996). mountains to breed (Melcher 1992). Breeding In northwestern British Columbia, Willow Ptar- site fidelity for yearling females appears lower migan were radio-marked on a breeding site and than for older birds as the majority of birds dis- followed to winter sites. Females moved an av- persed before or after their first breeding at- erage of 26-30 km (up to 88 km) and males 14- tempt. Thus ptarmigan, like a number of other 21 km (up to 90 km, Gruys 1993). Over winter, species, show reduced fidelity to breeding sites there were multiple movements between breed- in their first year of breeding (Black 1996). ing and winter sites. The search area was re- Data are limited on the consequencesof late stricted to 100 km around the breeding site and, spring dispersalto other breeding sites, however, because some birds were not located, dispersal dispersing females appeared to experience sim- distancesreported also were conservative. Given ilar reproductive successto those remaining on the problems of detecting individuals that dis- territories where they had previous breeding ex- perse the farthest, all empirical studies under- perience. Dispersal of females even in their pre- estimate spatial scalesof population connectivity laying period did not appear costly to survival. (Koenig et al. 1996, Lambrechts et al. 1999). Survival of radio-marked forest and arctic To understand linkages among spatially or- grouse was not related to distance moved, and ganized vertebrate populations, we need to ex- reproductive costs were not observed for birds amine the potential influence of seasonal move- switching territories or mates within sites (Beau- ments during post-breeding life history stages, dette and Keppie 1992, Hannon and Martin and winter environmental conditions on dispers- 1996, Smith 1997). Thus, ptarmigan females al patterns. When White-tailed Ptarmigan com- showed a propensity to disperse even after ini- plete breeding activities they abandon territories tiating breeding activities on a site, and did not and gather in late summer flocks at higher ele- incur survival or reproductive costs. Given that vations (Braun and Rogers 1971). Females with reproduction of White-tailed Ptarmigan shows broods normally go to the closest mountain top, strong age dependence, with older females rais- but males and unsuccessful females regularly ing more chicks (Martin 1995, Wiebe and Mar- move to nearby mountains (Herzog 1977, Braun tin 1998), even a small amount of breeding dis- et al. 1993, K. Martin, unpubl. data). In late fall, persal could make a significant contribution to birds are thought to move to winter sites in most local annual production. The ability to make sig- years (Braun et al. 1993). A study of 99 White- nificant movements without incurring survival tailed Ptarmigan in Colorado banded on winter or reproductive costs is an important pre-con- sites and relocated in summer reported average dition to the development of a dispersal rescue dispersal distances of 6.8 km for adult females, system. 8.5 km for yearling females, 2.5 km for adult males, and 4.0 km for yearling males (Hoffman SPATIAL SCALE AND INFLUENCES OF OTHER and Braun 1975). Birds moved in all directions LIFE HISTORY STAGES ON DISPERSAL but in a northwest predominance, with maxi- Data on seasonal movement patterns and dis- mum distances of 23 km for females and 10 km tances are required to determine the spatial scale for males. Hoffman and Braun (1975) predicted at which population exchange might occur. females dispersed within an area of 40-60 km2 These exist for few species. Median dispersal to breed. Seasonal movements from breeding to and recruitment distances for chicks banded on post-breeding and winter sites allow individuals breeding areas were 1.2 km for 126 males (up to assesspotential habitat for recruitment, and to 8 km) and 4.0 km for 40 females (up to 29 likely facilitate dispersal decisions. There is km) in earlier studies of White-tailed Ptarmigan much scope for research on starting and stop- at Rocky Mountain National Park and Mt. Evans ping rules to advance our understanding of dis- (Giesen and Braun 1993). These distances and persal behavior (Martin 1998). patterns underestimate true dispersal because Data on maximum movement distances such they relied on relocation of banded birds away as provided by Sutherland et al. (2000) and on potential dispersal barriers are required to iden- REGIONAL-LEVEL STOCHASTICITY tify the scale at which studies of multiple pop- Dispersal rescue is predicted to be most advan- ulations should be designed. Maximum move- tageous to population persistence when there is ment distances recorded for White-tailed Ptar- little regional stochasticity affecting reproduc- migan were for two males transplanted in spring tion, and least advantageouswhen regional sto- to another breeding site, that traveled 43 and 50 chasticity is high and/or when it affects adult km, respectively, acrossprimarily forested land- survival (Hanski 1991). We predict the heavy scape to return to their territories (Braun et al. reliance on external recruitment we observed for 1993). A successful transplantation of White- White-tailed Ptarmigan would persist only if re- tailed Ptarmigan to Pike’s Peak, an area of ap- gional-level stochastic events were relatively parently suitable but unoccupied habitat about rare (i.e., occur less frequently than the average 60 km from the nearest occupied habitat, sug- lifespan of individuals). The extreme weather gested this site exceeded normal ptarmigan dis- conditions in 1995 provided a type of natural persal distances (Hoffman and Giesen 1993; C. experiment on consequencesof regional failure E. Braun, pers. observ.). Demographic exchange on local population recruitment. Although sam- likely occurs between populations of White- ple sizes were small, recruitment in 1996 was tailed Ptarmigan within 5-10 km for males and lower than in previous years as expected, and 20-30 km for females. we were able to predict the severity of the im- We do not know whether our one-year telem- pact on population size among study sites. The etry study showing the surprising lack of migra- 1995 event had a relatively short-term impact on tion to winter sites was representative of over- population size given that most birds comprising winter movements on our study sites generally. 1996 populations were older and representedthe Although most of our radio-tagged birds did not most productive cohorts (Martin 1995, Wiebe leave their breeding sites over winter, there were and Martin 1998). By 1997, populations had re- normal numbers of birds resident on the tradi- covered. White-tailed Ptarmigan are relatively tional Guanella Pass winter site in 1997-1998 invulnerable to such regional-level events, especially if these do not greatly reduce adult sur- (K. Martin and C. E. Braun, pers. observ.). It is vival. Such regional-level catastrophes as de- possible that radio-tagging altered dispersal be- scribed for 1995 are not sufficiently frequent to havior, but we also observed high local recruit- threaten persistence of populations, given cur- ment of chicks in 1998 that were not radio- rent climatic conditions. However, climate tagged. In an earlier study, ptarmigan moved change is expected to result in increased fre- shorter distances in mild winters than in severe quency of stochastic weather patterns (Komer winters, and thus, weather patterns may influ- 1999), providing conditions that may result in ence the spatial scale of dispersal (Hoffman and increased demographic stochasticity and syn- Braun 1977). The 1997-1998 winter may have chrony in high-elevation habitats (Martin 2000). been sufficiently mild to allow exposed ridges White-tailed Ptarmigan may not fit all criteria where birds found adequate forage. Rock Ptar- for a traditional metapopulation structure be- migan (Lugopus mum) in Iceland showed an- cause they do not show regular extinctions and nual and regional variation in winter movements recolonizations of populations. However, many ranging from remaining on or near breeding of the principles and processesassociated with sites to migrating up to 300 km (0. K. Nielsen, metapopulations appear highly relevant to sta- pers. comm.). Even in sedentary specieslike the bility of ptarmigan populations. Ptarmigan exist Acorn Woodpecker (Melanerpes formicivorus), in spatially organized populations that are linked young birds made regular excursions from their by a well developed system of exchange, prin- natal area. When territorial vacancies occurred, cipally via natal dispersal. The exchange linking birds settled rapidly resulting in a discontinuous populations may not be frequent or reciprocal. pattern of recruitment (Koenig et al. 1996). The Given the demographic and environmental sto- role of offspring movements in facilitating sam- chasticity characteristic of this ground-nesting pling of potentially suitable recruitment loca- tetraonid, the well developed rescue pattern of tions is a fascinating and important area of pop- immigration of individuals from elsewhere in ulation connectivity research. the multi-population system appearsresponsible for maintaining stability in White-tailed Ptar- namics: concepts, models and observations.Biol. migan populations. J. Linn. Sot. 42:17-38. HANSKI,I. 1994. A practical model of metapopulation dynamics. J. Anim. Ecol. 63:151-162. ACKNOWLEDGMENTS HANSKI,I., AND M. GILPIN. 1991. Metapopulation dy- The study was funded by Natural Sciences and Engi- namics: brief history and conceptual domain. neering ResearchCouncil (NSERC) grants to KM and Biol. J. Linn. Sot. 42:3-16. by logistical assistancefrom the Colorado Division of HANSSON,L. 1991. Dispersal and connectivity in me- Wildlife (CEB). We thank the many able field assis- tapopulations.Biol. J. Linn. Sot. 42:89-103. tants who collected data on the ptarmiganproject, par- HARRISON,S. 1991. Local extinction in a metapopu- ticularly M. L. Commons who helped with radio-mark- lation context: an empirical evaluation. Biol. J. ing in 1997-1998. M. D. Mossop assistedwith man- Linn. Sot. 42:73-88. uscript preparation. We thank A. Caizergues, L. Elli- HARRISON,S., AND J. E QUINN. 1989. Correlated en- son, W. D. Koenig, M. M. Lambrechts, B. K. vironments and the persistence of metapopula- Sandercock,H. Steen, J. R. Walters, G. C. White, K. tions. Oikos 56:293-298. L. Wiebe, J. R. Young, and others for constructive HERZOG,I? W. 1977. Summer habitat use by White- comments during manuscriptcompletion. tailed Ptarmigan in southwestern Albeita. Can. Field-Nat. 91:367-371. LITERATURE CITED HOFFMAN,R. W., AND C. E. BRAUN.1975. Migration of a wintering population of White-tailed Ptarmi- ARTISS,T, AND K. MARTIN. 1995. Male vigilance in gan in Colorado. J. Wildl. Manage. 39:485-490. paired White-tailed Ptarmigan Lagopus leucurus: HOFFMAN,R. W., AND C. E. BRAUN. 1977. Character- mate guarding or predator detection? Anim. Be- istics of a wintering population of White-tailed hav. 49:1249-1258. Ptarmigan in Colorado. Wilson Bull. 89: 107-l 15. BEAUDETTE,P D., AND D. M. KEPPIE.1992. Survival HOFFMAN,R. W., ANDK. M. GIESEN.1983. Demogra- of dispersing Spruce Grouse. Can. J. Zool. 70: phy of an introduced population of White-tailed 693-697. Ptarmigan. Can. J. Zool. 61:1758-1764. BLACK, J. M. [ED.]. 1996. Partnershipsin birds, the KOENIG, WY D., D. VAN VUREN, AND F? N. HOOGE. study of monogamy. Oxford Univ. Press, Oxford. 1996. Detectability, philopatry and the distribu- BRAUN,C. E., R. W. HOFFMAN,AND G. E. ROGERS. tion of dispersal distances in vertebrates. Trends 1976. Wintering areas and winter ecology of Ecol. Evol. 11:514-517. White-tailed Ptarmigan in Colorado. Colorado K~RNER,C. 1999. Alpine plant life: functional plant Div. Wildl. Spec. Re:. 38, Denver, CO. ecology of high mountain ecosystems. Springer- BRAUN, C. E., K: MARTIN, AND L. A. ROBB. 1993. Verlag, Heidelburg, Germany. White-tailed Ptarmigan. In A. Poole. P. Stetten- LAMBRECHTS,M. M., J. BLONDEL,A. CAIZERGUES,P C. heim, and E Gill [ED;.], The birds of North Amer- DIAS, R. PRADEL,AND D. W. THOMAS.1999. Will ica, No. 68. The Academy of Natural Sciences, estimates of lifetime recruitment of breeding off- Philadelphia, and The American Ornithologists’ spring on small-scale study plots help us to quan- Union, Washington, DC. tify processesunderlying adaptation?Oikos 86: BRAUN,C. E., AND-G. E. ROGERS.1971. The White- 147-151. tailed Ptarmigan in Colorado. Colorado Div. LEWIS,E. 1904. The nesting habits of the White-tailed Game, Fish, &d Parks Tech. Publ. 27, Denver, Ptarmigan in Colorado. Bird Lore 6:117-121. co. MARTIN, K. 1995. Patterns and mechanisms of age- BROWN,J. H., AND A. KODRIC-BROWN.1977. Turnover dependentreproduction and survival in birds. Am. rates in insular biogeographv:immigration or ex- Zool. 35:123-141. tinction. Ecology 58:4i5:4449. - MARTIN,K. 1998. The role of animal behavior studies GIESEN. K. M.. AND C. E. BRAUN.1993. Natal disoersal I and recruitment of juvenile White-tailed Ptarmi- in wildlife science and management. Wildl. Sot. gan in Colorado. J. Wildl. Manage. 57:72-77. Bull. 26:91l-920. GREENWOOD,P J. 1980. Mating systems, philopatry MARTIN,K. M. 2000. Wildlife communities in alpine and dispersal in birds and mammals. Anim. Be- and subalpine habitats. In D. H. Johnson and T hav. 28: 1140-l 162. A. O’Neil [EDS.], Wildlife-habitat relationshipsin GRUYS,R. 1993. Autumn and winter movements and Oregon and Washington. Oregon State Univ. sexual segregation of Willow Ptarmigan. Arctic Press, Corvallis, OR, in press. 46:228-239. MARTIN,K., AND S. J. HANNON.1987. Natal philopatry HANNON,S. J., I. JONSSON,AND K. MARTIN. 1990. Pa- and recruitmentof Willow Ptarmiganin north cen- tagial tagging of juvenile Willow Ptarmigan. tral and northwesternCanada. Oecologia 71:518- Wildl. Sot. Bull. 18:116-118. 524. HANNON,S. J., AND K. MARTIN. 1996. Mate fidelity MARTIN, K., P B. STACEY,AND C. E. BRAUN. 1997. and divorce in ptarmigan:polygyny avoidanceon Demographic rescue and maintenanceof popula- the tundra, p. 192-210. In J. M. Black [ED.], Part- tion stability in grouse-beyond metapopulations. nershipsin birds, the study of monogamy. Oxford Wildl. Biol. 3:295-296. Univ. Press, Oxford. MELCHER,C. P 1992. Avifauna responsesto intensive HANSKI, I. 1991. Single species metapopulation dy- browsing by elk in Rocky Mountain National Park. M.Sc. thesis, Colorado State Univ., Fort upon local populationdynamics, p. 267-291. In 1. Collins, CO. Hanski and M. Gilpin [EDS.],Metapopulation bi- NICHOLS,J. D. 1996. Sourcesof variation in migratory ology: ecology, genetics and evolution. Academic movements of animal populations: statistical in- Press, New York. ference and a selective review of empirical results STACEY,l? B., AND M. TAPER. 1992. Environmental for birds, p. 147-197. In 0. E. Rhodes Jr., R. K. variation and the persistenceof small populations. Chesser, and M. H. Smith [EDS.], Population dy- Ecol. Appl. 2: 18-29. namics in ecological space and time. Univ. Chi- STENZEL,L. E., J. C. WARRINER,J. S. WARRINER,K. S. cago Press, Chicago. WILSON,E C. BIDSTRUP,AND G. W. PAGE. 1994. Long-distance breeding dispersal of Snowy Plo- PARADIS,E., S. R. BAILLIE,W J. SUTHERLAND,AND R. D. GREGORY.1999. Dispersal and spatial scale af- vers in western North America. J. Anim. Ecol. 63: 887-902. fect synchrony in spatial population dynamics. SUTHERLAND,G. D., A. S. HARESTAD,K. PRICE,AND K. Ecology Letters 2:114-120. E?LERTZMAN. 2000. Scaling of natal dispersaldis- SMITH, A. A. 1997. Dispersal and movements in a tances in terrestrial birds and mammals. Conserv. Swedish Willow Grouse Lagopus lagopus popu- Ecol., in press. lation. Wildl. Biol. 3:279. WIEBE, K. L., AND K. MARTIN. 1997. Effects of pre- SMITH,J. N. M., M. J. TAI?T,C. M. ROGERS,I? ARCESE, dation, body condition and temperature on incu- L. E KELLER,A. L. E. V. CASSIDY, AND W. M. bation rhythms of White-tailed Ptarmigan Lago- HOCHACHKA.1996. A metapopulationapproach to pus leucurus. Wildl. Biol. 3:219-227. the population biology of the Song Sparrow Me- WIEBE,K. L., ANDK. MARTIN. 1998. Age specific pat- Zosp& melodia. Ibisi38: 120-128. terns of reproduction in White-tailed and Willow STACEY.I? B.. V. A. JOHNSON.AND M. L. TAPER.1997. Ptarmigan Lagopus leucurus and L. lagopus. Ibis Migration within metapopulations: the impact 140:14-24.