The breeding biology of Rose-ringed Parakeets Psittacula krameri in England during a period of rapid population expansion

Abstract

Capsule The reproductive rate of Rose-ringed Parakeets in the UK was higher in 2001-2003 than previously estimated.

Aim To measure reproductive rate and the factors affecting this in Rose-ringed Parakeets in England during a period of rapid expansion in population of this potential invasive species.

Methods During 2001–2003, 108 nests were located and monitored in southwest London, southeast London and the Isle of Thanet. Nest survival using the Mayfield method, fledging success and the characteristics of the nest and surrounding vegetation were measured.

Results The mean date of first egg was 26 March ± 1.3 days and the median clutch size was 4 eggs. Mayfield nest survival rate was 72%. Reproductive success was 1.4 ± 0.3 young fledged per nest. Parakeets bred predominantly in ash (Fraxinus) and oak (Quercus). Nest trees had a diameter at breast height (DBH) of 73.7 ± 4.7 cm (mean ± se).

Conclusion Reproductive success in this parakeet was higher than previously estimated (0.8 young fledged per nest prior to 1999). This reproductive output is sufficient to explain the rapid increase in Rose-ringed Parakeet numbers (from an estimated 1500 birds in 1996 to nearly 10,000 birds by 2004) if Rose-ringed Parakeets have a low death rate similar to other parrots.

The Rose-ringed Parakeet is an established exotic in Britain (BOU 1983). The species was first reported breeding in Britain in 1969 (Hudson 1974, Lever 1977), and its population slowly increased to an estimated 500 birds in 1983 (BOU 1983) and 1500 birds by 1996 (Pithon & Dytham 1999b). Thereafter, the population rapidly increased to an estimated 5800 birds by 2001 (Butler 2002), with the bulk of the population located in the Greater London area and the Isle of Thanet, Kent. The rapid increase of this introduced species is a cause for concern from both an economic and conservation standpoint because it is a crop pest in its native range and the Rose-ringed Parakeet could potentially out-compete some native species for nesting cavities (Forshaw 1989, Juniper and Parr 1998, Strubbe & Matthysen 2007).

Two types of study on the breeding biology of birds are widely reported in the literature; nest-site selection and factors influencing reproductive success. Nest-site selection studies identify the habitat characteristics that a species requires to breed (Glue & Boswell 1994, Jones & Robertson 2001), while studies examining the factors influencing reproductive success predict fledging success on the basis of one or more variables such as age (Wiktander et al. 2001). Such studies are useful for evaluating the potential for spread of an invasive species (Hyman & Pruett-Jones 1995, Burger & Gochfield 2000), but to date, little work has been done on either nest-site selection or factors that influence reproductive success in Rose-ringed Parakeets in the UK (although see Newson et al 2011), although this has been addressed in Belgium (Strubbe & Matthysen 2007, 2009a, 2009b, 2011).

A study by Strubbe and Matthysen (2007) demonstrated that parakeet abundance is positively correlated with cavity abundance. Given the rapid increase of this species in England (Butler 2002, Holling & RBBP 2011, Baille et al. 2012) and Belgium (Strubbe & Matthysen 2007) it is worthwhile to identify additional habitat characteristics of areas where Rose-ringed Parakeets choose to nest, so that future researchers can predict the likely extent of the range of this species. In this paper we measure the reproductive rate of Rose-ringed Parakeets breeding in the UK during a period of rapid expansion (early 2000) in its numbers and range and describe the characteristics of nest sites and examine factors that influenced its reproductive success.

METHODS

Data collection

Nest searching was undertaken from mid-February 2001 to July 2003 in the Greater London area, around the coast of Kent, and in the town of Studland, Dorset, the three main areas where the species occurred at that time. Parks and other areas where public access was allowed were searched every other day from late February to July. Potential nests were located by examining cavities near perched parakeets as well as by following individual parakeets until one entered a cavity. In addition, 34 nestboxes were placed in six public areas on the Isle of Thanet during June and July of 2002. The nestboxes were based on the same measurements as those in a previous study in 1997 (Pithon and Dytham 1999a). As per the Pithon and Dytham (1999a) study, nestboxes were placed approximately eight metres above the ground.

The contents of the nest were examined using a miniature black-and-white video camera attached by a 20 m length of cable to a 4" LCD television made by Optimus (similar to the “burrow probe” reported elsewhere, e.g. Enkerlin-Hoeflich et al. 1999). A small ‘Maglite’ torch (Solitaire single cell AAA) was taped to the video camera to increase the illumination in the cavity. If the nest was less than 7.5 metres above the ground, a ladder was employed in order to reach the nest. Tree-climbing was used to reach nests that were located more than 7.5 m above the ground.

Nests were visited on a weekly basis, and the number of eggs and/or chicks was recorded. Inspection of the nests typically caused the adult(s) to leave, although some females remained throughout inspection. The adults returned shortly after nest inspection. Chicks were determined to have fledged if they were no longer in the nest after seven weeks, as long as there was no sign of depredation, because chicks in India are known to fledge after seven weeks (Lamba 1966). Nest survival was estimated using the Mayfield method (Mayfield 1975) with confidence limits calculated using Johnson (1979). We were unable to assess whether all nests in a given area were found as some nests may have failed before they could be found while other nests may not have been discovered.

In late June or July (after the young birds had fledged) a series of measurements was taken of the nesting tree. This included recording the tree species, the type of cavity (whether natural, woodpecker, or nestbox), the height of the tree (using a clinometer), the nest height (using a clinometer), orientation of cavity (using a compass), and DBH (diameter of the tree at breast height which is standardized as 1.4 m using a measuring tape). In addition, the tree species was determined and the DBH measured for all trees within 10m of the nest cavity (a tree was defined as a plant with a DBH of 3.0 cm or more).

Data analysis

Mean and standard error for the measurements of the nesting tree and environs were calculated, and a Rayleigh test was performed to examine whether the orientation of the opening of the nest cavity differed from random. A chi-squared test was performed to compare nest-site reuse among regions.

To examine the factors that affect reproductive success, a General Linear Model (GLM) for predicted clutch size was constructed using tree DBH, nest height, tree height, number of years that a nest had been occupied since the beginning of the study (coded as “0” if the nest was not known to have been occupied previously, “1” if it was occupied the previous year, and as “2” if the nest had been occupied for the two previous years), male age (based on the presence or absence of a pink neck ring), year, region, orientation of the nest cavity, type of cavity (natural, woodpecker, or nestbox), local tree stem count (measured by counting all trees within 10 m of the nest), local tree basal area (measured by summing the diameter of all trees within 10 m of the nest), and date of first egg (back-calculated if necessary from the hatching date). Some nests were used for multiple years and some trees contained multiple nests. In order to avoid pseudoreplication, a single nest for a single year was randomly selected from each tree (i.e. if a tree had three nests which were used for two years, only one nest for one year was used in the analysis).

A Mayfield logistic regression (Hazler 2004) was used in order to examine the effects of the above variables on overall nesting success. Mayfield estimates of fledging success (Mayfield 1961, 1975) were obtained as well as Johnson's estimate of standard error (Johnson 1979). The number of birds fledged was estimated by multiplying the Mayfield estimate of nest success by the average clutch size. All statistical analyses were performed using R 2.20. Data are presented as mean ± standard error except where noted.

RESULTS

Nest site choice

In total, 108 nests were located over the course of the study (33 in 2001, 37 in 2002, 38 in 2003), consisting of 46 nests on the Isle of Thanet (Kent), 34 in SE London, 26 in SW London and two nests at Studland (Dorset).

Nests were an average of 8.1 ± 3.8 m above ground and were in trees with a DBH of 73.7 ± 4.1 cm and a height of 19.5 ± 1.0 m. Parakeets nested in trees that had substantial shrub and tree cover surrounding them (13.1 ± 1.6 trees or shrubs within 10 m of the nest, combined basal area of shrubs and trees = 233.9 ± 21.0 cm). The orientation of the cavities in which parakeets bred did not differ from random (Rayleigh test, Z = 0.10, P = 0.41).

Parakeets nested in trees of 12 different genera, with Fraxinus accounting for 33% of the trees used (Table 1). Parakeet nests were observed primarily in old woodpecker nests (n = 58), but were also found in natural cavities (n = 31) and nestboxes (n = 19). Parakeets were not observed excavating their own cavities, and only one instance of enlarging an existing cavity (a Great Spotted Woodpecker Dendrocopos major hole) was noted. Nest-site re-use occurred most frequently on the Isle of Thanet, with 80% of nests reused during 2002, and 64.3% of nests in 2003; in contrast, nest reuse for the Greater London area was 47.6% during 2002 and 26.1% during 2003 but the difference between the two regions was not significant (χ² = 21.5, df = 23, P = 0.55). It is unknown whether nests that were reused in subsequent years were used by the same pair. Of the 34 nestboxes put up during the summer of 2002 on the Isle of Thanet, 23 were still present during the summer of 2003 and six (26.1%) were used by parakeets.

Table 1. A summary of tree species used as nest sites by Rose-ringed Parakeets.

Genus	Common Name	Used by parakeets
Acer	Maple	8 (11.6%)
Cedrus	Cedar	1 (1.4%)
Fagus	Beech	1 (1.4%)
Fraxinus	Ash	23 (33.3%)
Juglans	Walnut	1 (1.4%)
Platanus	Sycamore	4 (5.8%)
Populus	Poplar	6 (8.7%)
Quercus	Oak	15 (21.7%)
Robinia	Locust	1 (1.4%)
Salix	Willow	2 (2.9%)
Tilia	Lime	5 (7.2%)
Ulnus	Elm	2 (2.9%)
	Total	69 (100%)

Factors affecting reproductive success

Eggs were first recorded on 5 March, but may have been laid as early as 27 February (back-calculated from hatching date). A few parakeets laid eggs as late as mid-May but the mean date of first egg was 26 March ± 1.3 days. No second clutches were observed. Clutch sizes ranged from one to seven eggs, with a mean of 3.7 ± 1.2 eggs per clutch (median = 4, n = 77). Cavity type was the only variable to correlate significantly with clutch size, with nests in woodpecker cavities having significantly fewer eggs (β = -1.21, t = -2.8, P = 0.019; Table 2).

Table 2. A summary of the GLM predicting clutch size for the environmental variables.

Predictor variable	β ± se	t	P
Intercept	6.9 ± 1.3	5.3	<0.001
Local tree basal area	−0.001 ± 0.002	−0.5	0.62
Local tree stem count	0.016 ± 0.026	0.6	0.53
Tree height	−0.063 ± 0.033	−1.9	0.07
Orientation	0.002 ± 0.001	1.3	0.20
Nest height	0.053 ± 0.070	0.8	0.45
DBH	−0.002 ± 0.008	−0.3	0.79
Cavity – woodpecker	−1.2 ± 0.4	−2.8	0.008
Cavity – nestbox	−0.45 ± 0.53	−0.9	0.40
Occupied two years	0.37 ± 0.47	0.8	0.44
Occupied three years	0.60 ± 0.64	0.9	0.36
Subadult	−0.25 ± 0.55	−0.5	0.65
Southeastern London	0.19 ± 0.51	0.4	0.71
Studland	−1.8 ± 1.0	−1.8	0.08
Isle of Thanet, Kent	0.048 ± 0.62	0.1	0.94
Date of first egg	−0.018 ± 0.011	−1.6	0.12

In total, 87 nests fledged young while 19 failed, with a total exposure of 4166 days. The daily survival rate for each nest was 0.995 ± 0.001. The daily survival rate for the egg stage was 0.989 ± 0.003 and the daily survival rate for the nestling stage was 0.9989 ± 0.001. With an egg stage lasting 21 days and a nestling stage of 49 days on average 72% of nests successfully fledged. The average number of birds fledged was 1.4 ± 0.2 chicks. The Mayfield logistic regression revealed that none of the observed variables (type of cavity [natural, woodpecker, or nestbox], date of first egg, tree height, nest height, cavity orientation, DBH, or the extent of the surrounding understory vegetation) had a significant influence on chick or egg mortality rates (P > 0.1 for both regressions). Four instances of chick depredation were observed; in all instances this occurred when the chicks were less than two weeks old. All other instances of nest failure (both partial nest failure and complete nest failure) were due to unhatched eggs.

During this study, presumed sub-adult male Rose-ringed Parakeets (n = 15) were observed breeding, a result not reported elsewhere. The presumed sub-adult males (which lacked a rose-coloured ring) tended to have smaller clutches (3.3 ± 1.3 eggs) than did adult males (3.8 ± 1.2 eggs), although the differences in clutch size were not significant (t-test, t = 1.4, df = 14.8, P = 0.19). Similarly, there was no significant difference in Mayfield survival rates of nests of sub-adult males and adult males (7-test, t = -1.1, df = 14.1, P = 0.29).

DISCUSSION

Nest-site choice

As Rose-ringed Parakeets are secondary-cavity nesters, the presence of breeding pairs in an area will be dependent upon the number of cavities available. In this study, we found that most nests were in Quercus and Fraxinus spp. In the Netherlands, van Balen et al. (1982) found that forests dominated by Fagus and Quercus had an average of 6.2 cavities per ha. This suggests that cavities may be relatively common in woodlands with mature Quercus. Trees that grow rapidly tend to be more likely to have cavities (Ranius et al. 2009), which may explain why a relatively high proportion of nests were found in the fast-growing Fraxinus. Strubbe and Matthysen (2007) also found that parakeets were most common in areas dominated by Fraxinus. The DBH of trees used in this study was/were smaller than those trees used in northern India, which were 120 ± 6 cm (n = 15; Simwat and Sidhu 1973). Neither tree height nor amount of cover provided by nearby trees and shrubs have been reported in breeding studies of this species in its native range, and so no comparisons are possible.

Factors affecting reproductive success

The mean estimated date for the first egg (26 March) is similar to that reported in northern India (31° N, 11 March, Simwat and Sidhu 1973) and Bangladesh (24° N, 10 March, Hossain et al. 1993). Other authors have noted that some Psittaciformes tend to exhibit a high degree of nest initiation synchronicity (Renton and Salinas-Melgoza 1999), and a similar result was found in this study. Although second clutches have been reported in other populations (Hossain et al. 1993), no second clutches were found over the course of this study. Mean clutch size (3.7 ± 1.2 eggs) during this study was similar to the mean clutch size reported elsewhere (Lamba 1966, Simwat & Sidhu 1973, Shivanarayan et al. 1981, Hossain et al. 1993, Pithon & Dytham 1999a). The estimate of 1.4 young fledged per nest during 2001-2003 was comparable to that found in Bangladesh and southern India (Shivanarayan et al. 1981, Hossain et al. 1993), but considerably lower than that reported elsewhere in India (Lamba 1966). The estimate of 1.4 young fledged per nest during 2001-2003 was nearly twice the 0.8 birds fledged per nest (in the UK) reported previously (Pithon and Dytham 1999a). It is not clear whether reproductive output in the UK had increased during the years of our study or whether the low estimate by Pithon and Dytham (1999a) was a consequence of the small sample size (n = 12). It has been noted however, that reproductive output may be density-dependent in some introduced species, with low reproductive output at low densities (i.e. the Allee effect; see Lewis & Dareiva (1993)). It is possible that the increasing density of Rose-ringed Parakeets in England since the late 1990s may have had a positive effect on the number of young fledged.

Clutch size was smaller in woodpecker cavities than in natural cavities or nestboxes. This variation was unlikely to be due to varying predation because nest predation was only infrequently observed (n = 5), and in all instances was believed to have been caused by introduced Grey Squirrels Sciurus carolinensis which were observed entering the nest cavities. It is possible that this smaller clutch size may be related to increased interspecific competition for these cavities with species such as Green Woodpeckers Picus viridis. It is also possible that it may be related to cavity size, as Great Tits Parus major adjust clutch sizes in response to cavity size (Slagsvold and Amundsen 1992).

Human predation on nests (i.e. removing young from cavities) is a widespread problem for other parrot species (Thomsen & Mulliken 1991, Koenig 2001, Martuscelli 2003). However, it did not appear to be a widespread problem in the UK, with the exception of the Isle of Thanet. There, several trees had large notches cut into them. These notches were typically centred on old cavities, presumably to make it easier to remove nestlings. However, although human predation may occur on the Isle of Thanet, it did not do so during this study and fledging rates were similar to those in the Greater London area.

During this study, 6 of 23 (26%) nestboxes were used on the Isle of Thanet. In contrast, only 0.6% of nestboxes were used (one out of 175) in the Greater London area during 1998 (Pithon and Dytham 1999a). Nestbox design and positioning was similar in the two studies which suggests that suitable nesting sites may be a limiting factor for parakeets breeding at the Isle of Thanet, but not for parakeets breeding in the Greater London area.

Although it has been suggested that birds breeding in nestboxes may fledge more young than birds breeding in natural cavities (Nilsson 1986) other authors have not been able to detect a difference (Miller 2002). Over the course of this study, no differences in fledging rates were detected between parakeets breeding in natural cavities, woodpecker cavities, and nestboxes despite the fact that clutch sizes were significantly lower in woodpecker cavities.

The date of clutch initiation has been shown to influence reproductive success in a wide range of bird species including Great Tit Parus major (Barba et al. 1995), Spectacled Eider Somateria fischeri (Grand and Flint 1997) and Northern Flicker Colaptes aura (Ingold 1996), particularly if there is environmental stress such as cold temperatures or lack of food (Wiggins et al. 1994). However, other authors have found that the date of clutch initiation does not strongly influence reproductive success (e.g. Drever and Clark 2007). The timing of nest initiation did not affect fledging rates which may mean that Rose-ringed Parakeets in the UK do not suffer from a lack of food during the breeding season, perhaps due to access to bird feeders.

Many nests were discovered in Fraxinus (n = 23) and Quercus (n = 15). A survey of 243 Green Woodpecker nests in the UK found that the trees that this species typically chose were primarily oaks Quercus spp. and ash Fraxinus excelsior (Glue and Boswell 1994) and it is possible that parakeets preferentially nested in Fraxinus because many of these trees had existing woodpecker cavities.

The discovery that presumed sub-adult males are breeding in the UK was unexpected. However, without carefully examining Rose-ringed Parakeets in the hand, it is not possible to separate sub-adult males from adult females (Butler & Gosler 2004) and so it is possible that multiple females were attending the nests. However, an examination of the Rose-ringed Parakeet specimens at the Natural History Museum, Tring reveals that there is a sub-adult specimen with a slightly enlarged testis, which may indicate that sub-adult male Rose-ringed Parakeets occasionally breed.

Consequences of high reproductive success

It is unlikely that observed population growth of Rose-ringed Parakeets in the UK (from 1500 birds in 1996 to nearly 10,000 birds by 2004 and perhaps as many as 30,000 birds now; see Butler (2005) and Baillie et al. (2012) for more details) is entirely due to an increasing number of parakeets being released into the wild. Instead it is much more likely that the reproductive output observed during this study (1.4 ± 0.3 young fledged per nest) is driving the rapid population increase. For example, if survivorship is similar to the Puerto Rican Parrot Amazona vittata, i.e. 0.675 during year 1 and 0.848 for subsequent years (see Snyder et al. [1987] for more details), then the intrinsic rate of increase (r) is 0.27 ± 0.4. Consequently, a population of 1500 birds in 1996 should increase to 9400 – 18 000 parakeets by 2004 which is similar to the number reported in Butler (2005).

ACKNOWLEDGEMENTS

Thanks to Kristie Butler, David Griffin and David Ross for their assistance with this project. Thanks to Chris Feare, John Quinn, Tomasz Wesolowski, Alexa Lord, and two anonymous reviewers for their comments on an earlier draft. The analysis was carried out while supported by the University of Central Oklahoma. Will Cresswell was a Royal Society University Research Fellow during this study. Chris Redfern was the Editor for this submission.