J. Ornithol. 144, 69-80 (2003) Deutsche Ornithologen-Gesellschaft/Blackwell Verlag, Berlin ISSN 0021-8375
Non-territorial systems in corvids: the case for the Nutcracker (Nucifraga caryocatactes) in the Alps Antonio Rolando* and Loredana Carisio Dipartimento di Biologia Animale e dell'Uomo, via Accademia Albertina 17, 10123 Torino, Italy; Email:
[email protected] (*corresponding author)
Summary The Nutcracker is considered a territorial species even though some studies have shown that conflicts are rare. Previous research suggested that Nutcrackers from two Alpine sites did not use territories during autumn and winter. The aim of the present study was to describe the Nutcracker's spacing system during the reproductive period to provide a general interpretation of its spacing ecology. The study was carried out from March to May in the Alev~ forest (north-western Italian Alps) with radio-tracking and bird census techniques. Vocal activity was recorded throughout the year to make possible a better interpretation of the census results; recorded calls were also used for experimental play-back tests in the field. Home ranges were very small (median value of 22 ha, minimum convex polygon 95 %) and overlapped to a great extent (72 %, on average). No territorial confrontation was detected. Strangers were usually tolerated near the nest and a few experimental playback tests elicited no significant interaction initiatives. Present data indicate that the Nutcracker cannot be considered territorial, at least in our study area. We believe the spacing pattern of this species can be explained, at least partially, by its storing habit. In typical territorial systems, food resources are threatened by neighbours or strangers. In the Nutcracker's spring spacing system, food stores are not threatened because their location is known only to the owner. Thus, border trespassing is tolerated because strangers or neighbours are not food competitors. Estimates of Nutcracker density varied significantly among months, habitats and years. Arolla pine forests were preferred over the other habitats, probably because of autumn seed availability there. Vocal activity, which was weather-dependent, significantly changed among periods. Seasonal variations in vocal activity suggest caution when densities are compared among different periods.
Keywords: density estimate, home range, play-back, radio-tracking. Zusammenfassung Nicht-territoriale Systeme bei Corviden: das Beispiel des Tannenhiihers (Nucifraga caryocatactes) in den Alpen Der Tannenh~iher gilt als territorial, obwohl es nur sehr wenige Beobachtungen zu Territorialverhalten gibt. In der vorliegenden Untersuchung besch~iftigen wir uns mit dem r~iumlichen Verhalten yon Tannenh~ihem w~ihrend der Brutzeit. Dazu haben wir zwischen M~irz und Mai im Alev~-Wald (NW Italien) die Vrgel regelmggig gezghlt and telemetriert. Zudem wurde w~ihrend des ganzen Jahres die Rufaktivit~it aufgenommen. Weiterhin wurden Versuche mit Klangattrappen durchgeftihrt. Die Wohngebiete (home ranges) waxen sehr
U.S. Copyright Clearance Center Code Statement: 0021-8375/2003/14401-0069 $ 15.00/0
70
Journal ftir Ornithologie144, 2003 klein (Median 22 ha) und tiberlappten stark (durchschnittlich 72 %). Territoriale Auseinandersetzungen wurden nicht beobachtet. Fremde V6gel wurden am Nest meist toleriert und auf Klangattrappen wurde kaum reagiert. Diese Daten sprechen dafiir, dass der Tannenh~iher keine territoriale Art ist, zumindest im Untersuchungsgebiet. Das rfiumliche Verhalten l~isst sich wohl besser erklgren aus dem Versteckverhalten des Tannenh~ihers. W~ihrend bei territorialen Arten Nahrungsressourcen durch Nachbarn und Eindringlinge ,,bedroht" sind, gilt dies ftir die Nahrungsvorrfite des Tannenhfihers nicht, da diese nur dem individuellen Vogel bekannt sind. Deshalb kann er andere V6gel in seinem Streifgebiet tolerieren, da diese nicht als Futterkonkurrenten auftreten. Die Dichte an Tannenh~hern variierte stark zwischen Habitaten, von Monat zu Monat und zwischen Jahren. Im Untersuchungsgebiet wurden Zirbelkieferbestfinde bevorzugt, wohl wegen des herbstlichen Samenangebotes. Die Rufaktivit~it war witterungsabhfingig und variierte zeitlich. Diese saisonale Variation der Rufaktivit~it muss bei Bestandsaufnahmen nach rufenden V6geln beachtet werden.
Introduction One approach to the study of local spacing in animals is to consider two major spatial systems, i. e. home range and territory. A territory may be defined as "a more or less exclusive area defended by an individual or group" (Davies & Houston 1984). The single criterion of "defended area" is largely used in the literature, but those of "'exclusive area" and "sitespecific dominance" are also commonly employed (Maher & Lott 1995). However, the degree of overlap between adjacent territories is variable and the very concept of territorial exclusiveness is far from adequate. The home range may be defined as "that area traversed by the individual in its normal activities of food gathering, mating and caring for young" (Bun 1943). It should be mentioned that the home range is not all the area that an animal traverses, but rather the area in which it normally moves: thus occasional excursions outside its normal area should be ignored when the home range is defined (White & Garrott 1990). The home range concept has developed together with radio-telemetry studies focused on local movements and habitat use. When home range size is being estimated, the measured size increases very rapidly when fix-recording begins but a point of "sampling saturation" is reached as more and more positions are recorded. It is generally assumed that be-
yond this point the animal's range does not significantly increase (Kenward 1987). Home ranges of erratic individuals or floaters are obviously expected to go on increasing, but the existence of a true home range for these individuals can be questioned. However, since many studies have shown continuous range size increases even in apparently sedentary animals, the idea of a stable home range has recently been challenged (Gautestad & Mysterud 1995). Therefore, there are some problems with the concepts both of territory and of home range. Moreover, since it is often very difficult to ascertain the exclusive use of an area and to detect the keeping-out signals displayed, the borderline between these two spatial systems is not as clear-cut as may appear from the abovementioned definitions. The home range concept is used in a more general and comprehensive way than that of territory; indeed, spatial systems whose exclusiveness has not been ascertained are usually referred to as home ranges, even though they may instead be territories. These two terms are not true alternatives. It must be considered that in a strictly territorial species, the boundaries of the territory would also be those of the home range. On the other hand, in some species territories exist at the same time as home ranges: Chaffinches (Fringilla coelebs), for instance, use
A. Rolando & L. Carisio - Nutcracker territorial system
71
small territories included in much more extended home ranges (Maciejok etal. 1995). Moreover, home range may be considered an operational term used to describe the use of space, whereas that of territory is more appropriate for describing social systems.
exclusive area (i. e. no or little home range overlap).
Spacing patterns are frequently considered species-specific and, in fact, they are usually considered to be spatially consistent for each species (Brown & Orians 1970). Many corvid species are considered typically territorial (Coombs 1978, Goodwin 1986). A m o n g forest species, for instance, both the Jay (Garrulus glandarius) and the Nutcracker are considered territorial (Mattes 1978, Bossema 1979, Keve 1985, Andr6n 1990, Grahn 1990). Recent studies have challenged these firm beliefs: jays from an Italian coastal area did not display any territorial behaviour either in autumn-winter or in spring-summer (Rolando etal. 1995, Rolando 1998), and Nutcrackers from two Alpine sites did not use territories during autumn and winter (Rolando 1996, Rolando & Carisio 1999). However, it may be that Nutcrackers are territorially active only during the reproductive period, so movements and spacing systems should be studied mainly in spring. The present study was carried out during the breeding period, from March to May, in the Alev~ forest (north-western Italian Alps) by means of radio-tracking and bird census techniques. Vocal activity was recorded throughout the year to allow better interpretation of the census results; recorded calls were also used for experimental play-back tests in the field. The Alevb forest is mostly composed of Arolla Pines (Pinus cerebra) whose seeds are eaten or stored by birds starting from September. Spring data were compared with those previously collected in the same study area in autumn and winter (Rolando & Carisio 1999), so that it was possible to achieve a comprehensive analysis of the local spacing system of the Nutcracker throughout the year. In the present paper the term territoriality was operationally defined by use of the two criteria of defended (i. e. some form of interactive behaviour) and
Study area and methods Study area Our research was carried out from February to May 1997, 1998, 1999 and 2000 in the Alev6 forest (Varaita Valley, Cuneo province, NW Italian Alps). The study area (about 10 km 2) is covered by coniferous forests ranging from 1600 to 2300 m a.s.1. The Arolla Pine is the dominant tree, occurring in pure and mixed (with the Larch Larix decidua) coniferous forests. At low altitudes, common deciduous trees are the Ash (Fraxinus excelsior), Laburnum (Laburnum anagyroides) and Wych Elm (Ulmus montana). The understorey is formed by Bear's Grape (Arctostaphilos uva-ursi), Mountain Juniper (Juniperus communis), Rhododendron (Rhododendron ferrugineum) and Bilberry (Vaccinium myrtillus). The Alev~ forest grows on the northern slopes of the Varaita Valley, whereas the southern slopes are covered by larches only. Five habitat types were recognised for a study of habitat use: (1) Pure, thick Arolla Pine formations (PA), extending approximately from 1700 to 2100 m a.s,l.; (2) Pure, open Arolla Pine formations of high altitudes (PAH), extending from 2100 to 2300 m a.s.l., characterised by scattered trees growing in scree; (3) Mixed Arolla Pine-Larch formations (AL), extending from 1600 to 1900 m a.s.l., characterised by the dominance of Arolla pines over larches; (4) Pure Larch formations (PL); (5) Thin coniferous formations bordering meadows and pastures, from 1600 to 1700 m a.s.l., above the village of Castello (TF). Radio-tracking Nutcrackers were trapped with mist nets and aged according to Svensson's guide (1984). Nine birds were fitted with radio transmitters (2.6 g, i.e. less than 2 % of the birds' body mass), which were glued and tied to the base of one of the 2 central tail feathers (Kenward 1978). Individuals were tracked, at intervals of no less than 30 min, on 2-4 days each week and located by approaching the birds or by triangulation. A 200 x 200 m grid map was used. Ranges V software was used in the analysis of radio-tag data (R.E. Kenward & K.H. Hodder 1996, Institute of Terrestrial Ecology, Wareham, U.K.).
72 Range sizes were obtained through the minimum convex polygon method (hereafter MCP) considering 100 % and 95 % of fixes. Core areas were identified on the basis of 50 % contours. Differences between home range sizes were tested with nonparametric techniques (Kruskal-Wallis test; MannWhitney U-test) (Sokal & Rohlf 1995). The presence of stable home ranges was checked by two methods: 1) incremental area plots, by addition of fixes sequentially (birds with stable home ranges are expected to reach the maximum range size asymptotically, while floaters should increase their ranges progressively), and 2) analyses of weekly range auto-overlap, i. e. by calculation of a regression of the overlap between ranges of the same Nutcracker with the length of time elapsed between seven tracking days (Rolando & Carisio 1999). A Nutcracker showing a clearly decreasing auto-overlap with time would be classified as a floater. Territoriality was checked by range overlap analyses (95 % MCP), the Ranges V program producing a matrix of the percentage overlap of range A on B and B on A for any pair of ranges. Specific efforts were also made to detect interactions attributable to territorial disputes through searching for interactions at various distance bands around the nests. Intrusion rates In a strictly territorial system without helpers, extrapair individuals moving near the nest may be considered intruders. Helpers at the nest are not known for the Nutcracker, which is considered a territorial species. Thus the potential intrusion rate can be computed by monitoring extra-pair individuals inside the presumed territory. Some of our Nutcrackers regularly nested; in these cases, we were able to distinguish and record intruders at various distances from the nest simply by keeping under 6bservarion the untagged partner and radio-checking the tagged one. All the individuals seen or heard close to the nest (less than 10 m, 10-50 m, 50-100 m and over 100 m) were recorded during 30 min counts. Vocal activity and experimental playback tests Nutcracker calls were recorded with a Sony TC-D8 tape recorder and a Sony ECM 672 microphone. Preliminary spectrographic analyses (Avisoft-SASLab Pro 3.4b Software by Raimund Specht) showed a vocal repertoire made up of three major types of vocalisarions, i.e. a rasp call (usually composed of
Journal ftir Ornithologie 144, 2003 many successive units, i. e. many distinct "krrar"), a churr call and a mewing call (Cramp & Perrins 1994). Vocal activity was evaluated by sampling the number of calls of each vocalisation type per hour; counts were made in thick Arolla Pine formations, from March 1999 to February 2000. To avoid problems of pseudo-replication of data, differences in the number of calls/hour among periods and weather conditions were checked by repeated measures ANOVA using mean values (Sokal & Rohlf 1995).
A b u n d a n c e estimates Estimates of Nutcracker abundances were based on data obtained on transects along narrow tracks in the different habitats of the Alev6 forest; every bird seen or heard on a known length of transect was recorded. Estimates (number/km2) were obtained by consideration of the distances from the transect line within two belts. The number of birds detected declined with distance in an approximately exponential way (Rolando & Carisio 1999). Hence, the negative exponential model was used to calculate densities. According to this model, the probability of detection of a bird at x metres = e -~, where a is a constant. It can be shown that p = 1- e -aw where w is the centre to inner band distance. Hence a = (- ln(1 - p ) ) / w and the density (D) will be calculated with the formula D = 5aN/L, where N is the total number of birds seen or heard on a transect and L the transect length (Bibby etal. 1993). The width of the inner belt (w) was such that about half the records fell within it and half beyond. Accordingly, a different width was calculated for each habitat. Parametric tests were performed on density data after logarithmic transformarion. To avoid problems of pseudo-replication of data, density differences among habitats, months and years were checked by repeated measures ANOVA using mean values (Sokal & Rohlf 1995).
Results H o m e ranges Nine Nutcrackers were captured and radiotagged in the breeding period; four (i. e. 4, 5, 7 and 8) started the breeding cycle by building a nest and sitting on the eggs, but only one (4) successfully c o m p l e t e d it. H o m e range sizes calculated by M C R the n u m b e r o f fixes and
A. Rolando & L. Carisio. Nutcracker territorial system
73
Table 1. Age, number of fixes, number of tracking days, period of tracking and home range sizes for the tagged Nutcrackers. Home range sizes were calculated with the minimum convex polygon (MCP) analysis using 100 %, 95 % and 50 % of fixes respectively. Asterisks (*) indicate an individual which was captured in two consecutive years. A = adult, J = juvenile. Tab. 1. Alter, Anzahl Peilungen, Anzahl Telemetrie-Tage, Zeitraum der Telemetrie und WohngebietsgrOge (home range) von Tannenh~ihem. Die Wohngebietsgr6Be wurde nach der MCP-Methode (minimum convex polygon) ermittelt unter Berticksichtigung von 100 %, 95 % und 50 % der Peilungen. * = Vogel wurde in zwei aufeinanderfolgenden Jahren gefangen. A = Altvogek J = Jungvogel.
Nutcracker
age
N of fixes N of days tracked
1 2* 3
unknown A unknown
953 986 515
49 51 22
4* 5 6
A A unknown
993 302 122
37 13 6
7 8 9
A A J
442 387 310
21 19 26
the tracking period for each bird are shown in Table 1. Incremental area plots suggested stabilisation of the area with increasing number of fixes (added in chronological order) for all individuals. On average, the final stabilisation was achieved after 145 fixes, 40 and 300 fixes being extreme values. Analyses of weekly range auto-overlap showed that two Nutcrackers (7 and 9) had ranges whose auto-overlap significantly decreased with time, whereas all other did not. The individuals used very few 200 x 200 m squares (min. 4, max. 10, mean 6.6), so the ranges were very small, with a median value of 22 ha (MCP 95 %). Core areas were often restricted to a single square and, for breeding individuals, the nests were also located there. The smallest M C P core areas measured 4 ha (i. e. the single 200 x 200 m square). One adult male was captured and radio-tagged in 1997 (Nutcracker 2) and captured and tagged again in the subsequent year (4), when it successfully bred. Its home range remained virtually the same, with 66.5 % autooverlap. The range size of the sole juvenile
period of tracking March - May 1997 March - May 1997 April - May1997 March - May 1998 March - April 1999 April - May 1999 March - May 2000 March - May 2000 February - May 2000
homerange size (ha) MCP 100% 95% 5 0 % 42 50 28 24 24 16 16 16 98
40 32 28 22 14 14 16 14 78
4 4 14 4 4 4 4 8 8
captured during this study (Nutcracker 9) was the largest one (MCP 95 %, 78 ha). H o m e range overlaps for individuals captured at the same site in the same year (i. e. three birds in 1997, two in 1999 and three in 2000) are shown in Fig. 1. Ranges (MCP 95 %) overlapped to a great extent, with percentages of overlap ranging from 17.5 to 100 % (mean 71.7 % _+ 7.3). Core areas (MCP 5 0 % ) overlapped to a lesser extent (mean 43.4 % +_ 12.4) and did not overlap in three out of the seven possible comparisons between individuals. Intrusion rates The number of extra-pair individuals detected at various distances from the nest were counted for each breeding pair during 30 rain counts. A mean hypothetical intrusion rate (number of individuals/hour) was thus obtained. The cumulative number of extra-pair individuals detected during 290 hours of observations at nests are shown in Table 2. No terri-
74
Journal ftir Omithologie 144, 2003 1997
~ k e r
3
1999
Nutcracker 5
NuWracker 6 2OOO Nutcracker 9
2 °tN Fig. 1. The degree of overlap between ranges of Nutcrackers caught at the same site in 1997, 1999 and 2000. Overall home ranges (MCP 95 %) are given. Abb. 1. Grad der 0berlappung yon Wohngebieten (MCP 95 %) von Tannenh~hern, die am selben Ort 1997, 1999 und 2000 gefangen wurden.
torial confrontation was detected. Average hypothetical intrusion rates ranged from 1.8 (Nutcracker 7) to 7.2 individuals/hour (Nutcracker 5). Only 2.8 % of the intrusions recorded within 100 m of the nest were at less than 10 m, as compared with 36 % from 10 to 50 m and 61.2 % from 50 to 100 m. However, the expected propoItions, on the assumption that birds were randomly distributed in proportion to the area available (circle with a radius of 10 m, bands 10-50 and 50-100 m), were 1, 24 and 75 % respectively. Differences were significant (X2 = 63.15, p < 0.01). This suggests that breeding birds did not defend (or defended unsuccessfully) in the vicinity of the nest. A few playback tests, each carried out in the vicinity of the nests (20 metres), did not provoked any territorial confrontation. Abundance estimates Mean densities ranged from 8 to 373 birds/ kill 2. Apparently, the densities varied among habitats, months and years (repeated-measures ANOVA F4,24 = 3.3, p < 0.05; F2.26 = 5.3, p
