(Phthiraptera: Menoponidae, Philopteridae) on Feral Pigeons (Aves: Columbiformes: Columbidae) Terry D

712 Seasonal population dynamics of four species of chewing lice (Phthiraptera: Menoponidae, Philopteridae) on feral pigeons (Aves: Columbiformes: Columbidae) Terry D. Galloway,1 Robert J. Lamb

Abstract—Seasonal dynamics of louse (Phthiraptera) populations on feral pigeons, Columba livia Gmelin (Aves: Columbiformes: Columbidae) were investigated from 2003 to 2012 in southern Manitoba, Canada. Pigeons were infested with: Philopteridae – Campanulotes compar (Burmeister), Columbicola columbae (Linnaeus), and Coloceras tovornikae Tendeiro; Menoponidae – Hohorstiella lata (Piaget). We consider the hypothesis that four species living on the same host show similar seasonal dynamics, coordinated by the life history of the host. Adults of both sexes and nymphs of all four species were present on pigeons throughout the year, consistent with continuous feeding and reproduction. Campanulotes compar and C. columbae populations were low in spring and peaked in September, with C. columbae showing greater seasonal changes for all population parameters. Coloceras tovornikae showed two annual peaks in abundance in spring and late summer, and H. lata was most abundant in the cold months of the year. Over 10 years, the four species showed distinct seasonal dynamics, although they live on the same birds. Seasonal patterns provided no evidence that louse reproduction or abundance is coordinated by the long breeding and moulting seasons of the host.

Introduction (Galloway and Lamb 2014). Campanulotes compar and C. columbae are abundant and their populations Four species of chewing lice (Phthiraptera) are relatively stable from year to year, whereas infest feral pigeons, Columba livia Gmelin (Aves: C. tovornikae and H. lata are less abundant and less Columbiformes: Columbidae) in Manitoba, stable than the former species. Canada: Philopteridae – Campanulotes compar The habitat of bird lice, the surface of the body (Burmeister), Columbicola columbae (Linnaeus), of the host, has a relatively constant temperature and Coloceras tovornikae Tendeiro; Menoponidae – through the season, in comparison with the Hohorstiella lata (Piaget) (Galloway and Palma changes in ambient air temperature experienced 2008; Galloway and Lamb 2014). These species by the host, particularly in a continental climate occupy the skin or feathers of their host (Nelson and such as that of Manitoba. Seasonal changes in Murray 1971; Singh et al. 2000), often on the same louse populations might be expected to be more bird, and all have chewing mouthparts: three are muted than those of multivoltine insects more feather feeders (Philopteridae) and one feeds on openly exposed to ambient temperatures (Woodman blood (Menoponidae) (Galloway and Palma 2008). and Dicke 1954). On the other hand, many bird Coloceras tovornikae appears to be a relatively recent species exhibit seasonal changes in their life introduction into North America (Galloway and histories associated with migration, moulting, and Palma 2008), compared with the other three species, reproduction, which may affect the seasonal which may have been introduced to North America biology of their ectoparasites. Foster (1969) along with the rock pigeon. Nevertheless, each concluded that the life cycles of a warbler and its species has a characteristic population biology blood-feeding chewing lice (Ricinus picturatus

Received 19 May 2014. Accepted 24 September 2014. First published online 20 January 2015.

T.D. Galloway,1 R.J. Lamb, Department of Entomology, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2

1Corresponding author (e-mail: [email protected]). Subject editor: Heather Proctor doi:10.4039/tce.2014.84

(Carriker) and Menacanthus Neumann species) Prairie Wildlife Rehabilitation Centre (Winnipeg, were synchronised: oviposition by lice coincided Manitoba, Canada). All came from southern with the short nesting period of its host. She Manitoba, Canada, mostly from Winnipeg. The found no evidence that oviposition occurred when majority of birds were euthanised soon after birds moulted. Hamstra and Badyaev (2009) arrival, immediately placed in sealed plastic bags, hypothesised that ectoparasite populations thrive and frozen for at least 48 hours until they could be at particular points in the life history of their avian processed. Lice were collected by washing hosts, because the energetic requirements of pigeons (Mironov and Galloway 2002), and this moulting or breeding limit immune responses and method removes nearly all lice from the host preening behaviour, the main defences against (Clayton and Drown 2001; T.D.G., personal ectoparasites. Little information is available on observation). The samples of pigeons and lice are the seasonal dynamics of lice on birds to assess described in detail by Galloway and Lamb (2014). these hypotheses. To date, studies of seasonal For the years 2003–2012 the following data dynamics are based on examination of museum were considered for each species of louse and specimens for the presence of eggs (Foster 1969) each pigeon: collection date, number of lice, or collections from birds through one year (e.g., number of adult females and males, and number Boyd 1951; Woodman and Dicke 1954; Kettle of nymphs. Nymphs were not sexed. Eggs were 1983; Spitznagel 1985; Chandra et al. 1988; not removed reliably by washing and so were not Petryszak et al. 2000; Hamstra and Badyaev tallied. The 10 chicks in the sample were excluded 2009), or over a relatively short period of time, from parameter estimates because they came up to three years (Boyd 1951; Bergstrand and from a restricted part of the season, whereas Klimstra 1964; Tuleshkov 1965; Baum 1968; the 49 juveniles and larger sample of adults Eveleigh and Threlfall 1976). were distributed throughout the year. Data for Feral pigeons in Winnipeg, Manitoba are 11 pigeons were excluded because they were abundant and non-migratory, with high winter considered outliers with extremely high infesta- counts compared to many other locations in North tions of one louse species (>10 times mean American (Taylor 2003). Pigeons alternate intensity) (Galloway and Lamb 2014). If they had between phases of roosting communally and a been included, the high louse numbers would long breeding season. In Manitoba, pigeon eggs have biased estimates of peak seasonal abun- have been observed in nests between 16 April and dance, assuring that the month of peak seasonal 29 July, and at least two broods may be produced abundance was determined by the collection date each year (Taylor 2003). In one previous study of of a single, heavily infested bird. The excluded seasonality of louse populations on pigeons in birds comprised 2% of the total, leaving 542 birds. Poland, Petryszak et al. (2000) found the greatest The number of pigeons available in each year abundance of C. compar in July and C. columbae ranged from 26 to 126, and in seven of the years in autumn. 41 or more birds were assessed each year. We investigate the seasonal population Data for the 10 years were ﬁrst partitioned into dynamics of four species of lice on feral pigeons samples for each month, providing a mean of to assess whether differences in seasonality 45 ± 9 (SD) pigeons per month. Abundance, contribute to the species-speciﬁcity of population prevalence, intensity, ratio of males to females, processes of these lice and whether louse season- and ratio of nymphs to females were estimated for ality is coordinated by pigeon life history. each month and louse species. Prevalence is the proportion of birds infested by a louse species, intensity is the number of lice of that species on an Materials and methods infested pigeon, and louse abundance equals prevalence multiplied by intensity (Rózsa et al. Ten years of data are used to describe seasonal 2000). This monthly partition provided an overall patterns in the abundance of lice on pigeons picture of the dynamics of each louse species over salvaged from rehabilitation hospitals at the the year, but no indication of how uniform any Wildlife Haven (Manitoba Wildlife Rehabilitation seasonal patterns were from year to year. In some Organization, Winnipeg, Manitoba, Canada) and years, the number of pigeons available some

© 2015 Entomological Society of Canada 714 Can. Entomol. Vol. 147, 2015 months was inadequate to effectively compare the same month or season ranked lowest or monthly estimates among years, in part because of highest was determined. If the same month or the large difference in louse abundance among season always, or nearly always, ranked lowest pigeons. Many pigeons had no lice, particularly or highest, this result provided evidence for the less common louse species, assuring that seasonal patterns in parameter values. abundance was not normally distributed (Galloway The degree of seasonality, or degree to which a and Lamb 2014). parameter value changed over the year, was To help assess year-to-year variation in season- quantified for each louse species in two ways. ality, data also were partitioned into four seasons: First, the coefficient of variation (CV) and its 95% winter (December, January, and February), spring confidence interval (corrected following Sokal (March, April, and May), summer (June, July, and and Rohlf (1981)) was calculated for the August), and autumn (September, October, and 12 monthly estimates. The confidence intervals November). For each of seven years of the study allowed apparent differences among louse species (2004, 2005, 2007, 2009, 2010, 2011, and 2012), to be evaluated. Coefficient of variation included this partition provided 41–126 pigeons per year, or variation that resulted from seasonal changes and usually, 10 or more pigeons per season: mean of also variation due to “error” (imprecision in the 16 ± 12 (SD, n = 28, range 4–66) birds per season. monthly estimates). The error was likely to be The other three years had 26–33 pigeons per year, higher when the number of pigeons with lice and too few to provide useful estimates for a season. the numbers of lice per pigeon were low (e.g., The number of pigeons collected in a season C. tovornikae and H. lata), which might have represents the sample size for abundance and inflated the estimate of seasonality. To reduce this prevalence, but for the other three parameters possible source of bias, the degree of seasonality sample size is the number of pigeons with lice was also quantified as the range for the highest (prevalence multiplied by the number of pigeons). and lowest monthly estimate, divided by the When prevalence was low, the sample size of birds annual mean for all months, eliminating variation with lice was sometimes small increasing variance from month to month and therefore much of the in estimates substantially (see Results). error. The disadvantage of this approach was that The uniformity of seasonal patterns for the no confidence intervals could be estimated. The population parameters was assessed in a number similarity in the estimates resulting from the two of ways. First, two-way analyses of variance were methods (CV versus range adjusted for the mean) conducted for each parameter, attributing varia- was compared for each parameter and the four bility to year and month, or year and season louse species, using Pearson’s product moment (SYSTAT 2010). A significant F-value for month correlations (SYSTAT 2010). The two methods or season confirmed that a parameter varied over were highly correlated (r > 0.98, P < 0.02, the year. Because parameter estimates often were n = 12 months), and so CV and its confidence not normally distributed and their variances limits were adopted as an index of the degree of were sometimes not homogeneous, the analyses seasonality. To further validate apparent differ- were repeated with a Kruskal–Wallis one-way ences among species, the process of estimating analysis of ranks or a Mann–Whitney rank sum the degree of seasonality was repeated using test (SYSTAT 2010). These non-parametric tests estimates for winter, spring, summer, and autumn led to essentially the same conclusions as analyses (correlation of CV versus range adjusted for the of variance. Patterns of abundance were compared mean: r = 1.0, P < 0.0001, n = 4 seasons), which between the first and second five years of the provided a less precise characterisation of seaso- study, to assess whether they were consistent. nal change (four seasons instead of 12 months) Pearson’s product moment correlations (SYSTAT but larger samples of pigeons for each estimate. 2010) were used to compare monthly abundances between the two five-year intervals, and also to assess the similarity of the seasonal patterns Results between species of louse. Finally, the monthly and seasonal parameter values were ranked for each In every month of the year, adults and/or year, and the number of years (out of seven) that nymphs of each of the four species of lice were

Fig. 1. Mean monthly abundance (n = 10 years) for study occurred between 5 April and 21 October. four species of lice collected from feral pigeons in Five chicks had lice, mostly C. compar (mean of southern Manitoba, Canada. The pigeons most heavily 47 lice per bird), and 93% of the 609 lice of all infested with lice (2%) are outliers and excluded (see four species came from three chicks collected in text). The monthly sample sizes of pigeons are given September and October. The 50 juveniles (birds on the upper x-axis. that had fledged but retained some juvenile plumage) were collected in every month except March, the last winter juvenile was collected on 27 February and the first spring juvenile on 27 April, leaving a two-month gap. Juvenile birds occurred most frequently through the summer and autumn. The prevalence of juveniles with lice for the four species were C. compar – 72%, C. columbae – 72%, C. tovornikae – 30%, H. lata – 20%, respectively, higher than for chicks. Similar numbers of C. compar were found on chicks and juveniles, but other species of lice had more than twice as many lice on juveniles as on chicks. Infestation of juveniles was similar to that of adults (Galloway and Lamb 2014). The seasonal patterns of louse abundance on pigeons were based on mean monthly estimates over 10 years. These patterns might reflect real seasonal changes or random variation in abundance from month to month. Partitioning the variance in abundance of lice (square root transformed) between year and month for individual pigeons, revealed differences among months for all four species (F = 2.7, P < 0.002, df = 11, 438), consistent with seasonal patterns. However, abundance was not normally distributed for any of the four species and variance in abundance present on some pigeons (Fig. 1). The two differed among months even after transformation, most common louse species, C. compar and casting doubt on the test. Partitioning variance of C. columbae, had one annual peak in abundance mean monthly abundance between year and in September with a mean of about 100 lice per month (pooling data from all pigeons collected bird. Both species had relatively low abundance in a month) eliminated the unequal variances, from February through May. In contrast, abundance but means were not normally distributed even if of the two less common lice, C. tovornikae and log-transformed. Differences were detected H. lata, was usually an order of magnitude lower among months for C. compar and C. columbae and showed more complex seasonal patterns. (F = 2.7, P < 0.005, df = 11, 104 and F = 7.8, Both peaked in March at about six lice per bird. P < 0.001, df = 11, 104, respectively), but not for Coloceras tovornikae had a second peak from C. tovornikae or H. lata (F = 1.4, P = 0.208, July to September, and about one louse per bird df = 11, 104 and F = 1.6, P = 0.122, df = 11, 104, from October through December. Hohorstiella respectively). Partitioning data into four lata had less than one louse per bird from April to seasons (pooling data from all pigeons collected October before abundance began to rise in in a season) gave essentially the same result (not November. shown), regardless of whether data were log- Feral pigeons have a long breeding season in transformed. Kruskal–Wallis one-way analyses southern Manitoba. The 10 chicks (birds without on ranks (which does not assume data are fully developed flight feathers) collected in this normally distributed) confirmed differences

Table 1. Consistency in seasonal patterns of abundance for Coloceras tovornikae and Hohorstiella lata between the ﬁrst (2003–2007) and second (2008–2012) ﬁve years of a study on lice on feral pigeons in southern Manitoba, Canada.

Abundance (2003–2007) Abundance (2008–2012)

Months Mean Median n Mean Median n Coloceras tovornikae 1, 10, 11, 12 1.22 0.50 17 0.75 0.33 18 2, 4, 5, 6 2.58 1.92 19 3.98 1.25 19 3, 7, 8, 9 3.82 3.33 13 4.05 2.00 19 Kruskal–Wallis one-way analysis on ranks (2 df) P = 0.131 P = 0.048 Hohorstiella lata 4, 5, 6, 7, 8, 9 0.77 0 24 0.31 0 29 1, 2, 3, 10, 11, 12 1.26 0.25 25 3.97 1.00 27 Mann–Whitney rank sum test P = 0.225 P = 0.01 For C. tovornikae, month 3 was included with 7, 8, and 9 because from 2003 to 20012, abundance was highest in these four months. among months for C. compar and C. columbae differences in the patterns occurred. For C. compar, (H = 23.4, P = 0.016, df = 11 and H = 55.8, peak abundance occurred in January, June, and P < 0.001, df = 11, respectively), but not for November in individual years, but seven of C. tovornikae and H. lata (H = 16.0, P = 0.142, 10 peaks occurred in August–October. Seasonal df = 11, and H = 13.3, P = 0.275, df = 11, minima for this species usually occurred from respectively). February to May, but also in August and December For both C. compar and C. columbae, which had in individual years. For C. columbae, peak abun- the clearest seasonal patterns, abundance declined dance always occurred from July to October, in autumn, was lowest in winter and spring, and usually in September. Seasonal minima occurred increased in summer to September (Fig. 1). from January to June. Monthly average abundance was correlated for the Possible seasonal patterns for C. tovornikae and two species (Pearson’s product moment correlation: H. lata were not confirmed by analysing abundance r = 0.865, P < 0.001, n = 12), confirming the for individual months because of the small numbers similarity in their seasonal patterns. The patterns of infested pigeons. However, abundance of were not identical, however, with C. compar having C. tovornikae fromFebruarytoSeptemberwas a mean monthly abundance over 10 years of 25.4 higher than from October to January (Mann– lice per bird from February through May in com- Whitney rank sum test: T = 1431, P = 0.004, parison with 2.8 for C. columbae (Mann–Whitney n = 70, 35). Abundance was low in early winter, rank sum test: T = 1642, P < 0.001, n = 36, 36). intermediate in February and spring, and high For both species, seasonal patterns were consistent in March and summer, consistently so in both between the first (2003–2007) and second (2008– five-year periods (Table 1). The spring peak 2012) five years of the study with monthly abun- occurred from February to June, most frequently in dances in the two periods correlated for C. compar May, whereas the summer peak occurred from July (Pearson’s product moment correlation: r = 0.547, to November, with 4.5 months between peaks, on P = 0.066, n = 12) and for C. columbae average. Abundance of H. lata was higher in the (r = 0.839, P < 0.001, n = 12), and with the colder part of the year (October–March) than in patterns correlated between species in both time the warmer months (April–September) (Mann– periods (2003–2007: r = 0.855, P < 0.001, n = 12; Whitney rank sum test: T = 3150, P = 0.007, 2008–2012: r = 0.662, P = 0.019, n = 12). n = 52, 53), consistently so in both five-year peri- Although seasonal patterns were clear for mean ods (Table 1). However, H. lata could be absent monthly abundance over 10 years, year-to-year from the sampled birds in any month of the year.

Peak abundance occurred in cold months seven of Fig. 2. Mean monthly prevalence and intensity of 10 years and in warm months three of 10 years. For infestation (n = 10 years) for Campanulotes compar both C. tovornikae and H. lata, patterns of seasonal and Columbicola columbae on feral pigeons from abundance were more clearly expressed from 2008 southern Manitoba, Canada. The monthly sample sizes to 2012 than from 2003 to 2007 (Table 1). of pigeons are as shown in Fig. 1. Pigeons with the highest intensity of lice (2% most infested), which were excluded from estimates of monthly abundance, tended to follow the seasonal patterns for each louse species. For C. compar, the most heavily infested pigeons were found from June to October, when peak abundance typically occurred. The same was true for C. columbae, with the most infested birds found from July to September, over a narrower range of months than C. compar, as was the case for peak abundance (see above). For C. tovornikae, however, the highly infested pigeons were found in July and November when monthly abundance was not at its highest. For H. lata the most heavily infested pigeons were found in March when abundance peaked and in October when abundance was beginning to rise as winter approached. From April or May to September, the abundance of C. compar increased threefold (Fig. 1) due in part to a small increase in prevalence, but due primarily to a 2.5-fold increase in intensity (Fig. 2). In the same interval, C. columbae increased 21-fold (Fig. 1), due to a threefold increase in prevalence and a sevenfold increase in intensity (Fig. 2). Prevalence varied less over the season than intensity, and both species had high prevalence in winter but relatively low prevalence in April and May. was highest in the six coldest months and intensity Patterns in these two components of abundance also peaked in those months (Fig. 3), consistent were confirmed in one-way analyses of variance on with abundance (Fig. 1). Although no seasonal ranks of seasonal mean values for the seven years differences were detected, prevalence was highest with largest collections of pigeons: prevalence and in winter and autumn and intensity was highest in intensity were lowest in spring (Table 2). For winter and spring, consistent with the relatively C. tovornikae, prevalence showed a March and a high abundance of this species in cold months late summer peak (Fig. 3), consistent with the two (Table 2). peaks in abundance (Fig. 1), but when data were Adult lice of both sexes were present on some partitioned into four seasons, prevalence was feral pigeons each month of the year. Sex ratio uniform (Table 2). For C. tovornikae, intensity was was near 1:1 and relatively uniform from season erratic but low from October to February (Fig. 3), to season for C. compar and C. columbae consistent with the low value for intensity in winter (Table 2). Males were sparse in winter for (Table 2). The relatively smooth curve for pre- C. tovornikae, consistent with the low intensity of valence, contrasted with the curve for intensity this species in that season (Table 2), and perhaps (Fig. 3), probably reflects the differences in sample reflecting higher death rates of males in winter. sizes for the two measures: all birds contribute to The ratio of males to females was < 0.4 for H. lata the sample size for prevalence but only about 30%, throughout the year, lower than for any of the on average, the birds with lice, contribute to the other species, with no seasonal pattern in the sex sample size for intensity. For H. lata, prevalence ratio detected (Table 2).

Table 2. Seasonal estimates of population parameters (prevalence, intensity, ratios of males and nymphs to females) for lice on feral pigeons collected over seven years in southern Manitoba, Canada.

Season* Campanulotes compar Columbicola columbae Coloceras tovornikae Hohorstiella lata Prevalence Winter 0.84 0.90 0.31 0.25 Spring 0.74 0.34 0.31 0.19 Summer 0.87 0.64 0.35 0.10 Autumn 0.92 0.91 0.32 0.28 P† 0.037 < 0.001 0.881 0.141 Intensity Winter 53.2 29.4 3.6 6.7 Spring 36.4 8.1 9.0 8.8 Summer 78.3 76.7 7.4 4.1 Autumn 76.2 108.0 6.6 3.3 † P 0.056 < 0.001 0.190 0.786 Ratio of males to female Winter 1.00 1.02 0.48 0.24 Spring 1.12 0.93 0.55 0.34 Summer 1.16 1.00 0.66 0.25 Autumn 1.21 0.96 1.34 0.39 P† 0.086 0.205 0.048 0.972 Ratio of nymphs to female Winter 3.02 3.86 2.09 1.72 Spring 2.85 4.38 1.71 1.69 Summer 3.34 7.52 2.14 1.57 Autumn 3.59 5.65 1.55 1.11 P† 0.430 0.010 0.824 0.965 * n = 119, 117, 123, and 97 pigeons assessed in winter, spring, summer, and autumn, respectively. Sample sizes for intensity and the two ratios = prevalence times n. † Probability (P) that seasonal estimates differ for each parameter and species based on Kruskal–Wallis one-way analysis of variance on ranks (3 df) (SYSTAT 2010), estimated after eliminating one to ﬁve pigeons (1%) with the highest ectoparasite intensity.

Nymphs were present on some pigeons every winter than in summer. For H. lata, prevalence month of the year for all four species, and always was usually (six of seven years) lower in summer more abundant than females (Table 2). No than in autumn. differences in the ratio of nymphs to females was The degree of seasonality differed among the detected among seasons for three of the species, four species (Table 3). Seasonality was quantified but for C. columbae the proportion of nymphs was as a coefficient of variation for monthly means highest in summer and autumn when intensity (10 years) for parameters. These means were not peaked for that species (Table 2). always normally distributed; therefore, coefficient The seasonal patterns were usually consistent of variation should be considered an index of the from year to year over seven years. For example, degree of seasonality, with confidence intervals prevalence of C. compar was always highest in approximate estimates. Campanulotes compar autumn and usually lowest in spring (five of seven showed the least seasonality for all parameters, years). Intensity was usually lowest in spring (five with CV usually less than half that for C. columbae of seven years). For C. columbae, overall abun- and H. lata (Table 3). Coloceras tovornikae tended dance was always highest in autumn and lowest in to have an intermediate degree of seasonality, spring. Prevalence was always and intensity was except for sex ratio, but is perhaps a special case usually (five of seven years) lowest in spring. The because it exhibited two annual peaks in abundance ratio of nymphs to females was always lower in rather than one.

Discussion winter and in summer. The fourth species, H. lata, has its highest abundance in the coldest months of The two more common species of lice that the year. These seasonal patterns are consistent infest feral pigeons in Manitoba, C. compar between two five-year periods, although the and C. columbae, have one late summer peak month of peak abundance varies from year to in abundance, while the less common species, year, being less consistent for the uncommon C. tovornikae, has two peaks in abundance, in late species than the common ones. Even though pigeon populations were assessed over a longer Fig. 3. period (10 years) and the sample of birds was Mean monthly prevalence and intensity of fi infestation (n = 10 years) for Coloceras tovornikae larger than in previous studies, con rming that and Hohorstiella lata on feral pigeons from southern these seasonal patterns are real was challenging. Manitoba, Canada. The monthly sample sizes of Month-to-month changes in abundance or the pigeons are as shown in Fig. 1. components of abundance and year-to-year changes in the timing of the seasonal dynamics require large monthly samples of the host, particularly if a louse species is present on a small proportion of birds in the population. The aggre- gated distribution of lice on birds contributes to this problem. Nevertheless, sufficient data were collected for the louse populations on pigeons in Manitoba to show that the four species have distinct seasonal population dynamics. The one previous study of seasonality of louse populations for pigeons, in Poland, reports the highest abundance of C. compar in July and C. columbae in autumn (Petryszak et al. 2000). This one-year study is consistent with the seasonal patterns observed in Manitoba some years, but not others. It is too short to assess whether these species have the same seasonal dynamics in Poland and Manitoba. Petryszak et al. (2000) collected lice by ruffling the feathers of their hosts, so mean intensities of infestation by the various species of lice were likely underestimated (Clayton and Drown 2001; T.D.G., personal observations). All four species of lice are present on pigeons throughout the year, with males, females, and nymphs present in every month of the year.

Table 3. Index of seasonality (CV* ± 95% conﬁdence interval) for population parameters (abundance, prevalence, intensity, ratios of males and nymphs to females) for monthly estimates of lice on feral pigeons collected over 10 years in southern Manitoba, Canada.

Parameter Campanulotes compar Columbicola columbae Coloceras tovornikae Hohorstiella lata Abundance 40 ± 18 90 ± 42 69 ± 32 101 ± 49 Prevalence 11 ± 534± 15 31 ± 14 52 ± 24 Intensity 36 ± 17 78 ± 36 51 ± 23 74 ± 34 ♂/♀ 8 ± 418± 837± 17 79 ± 36 Nymphs/♀ 24 ± 11 34 ± 16 31 ± 14 46 ±21 * Coefﬁcient of variation (%) measures the amount of variation among 12 monthly estimates, for a mean of 45 (range 31–69) pigeons per month.

The ongoing presence of all stages suggests that None of the louse species shows a change in development and reproduction are continuous abundance that matches the timing of reproduc- throughout the year, although that conclusion tion by its host. Campanulotes compar and needs to be confirmed by establishing that eggs C. columbae are near their population minima are laid continuously. An alternative hypothesis is through winter until a few months past the onset that nymphs take months to reach adulthood, or of breeding by the host. Similarly, the patterns of perhaps females enter a reproductive diapause. seasonal dynamics for C. tovornikae (peaking first These explanations seem unlikely, given that the lice before pigeons breed in spring and again four live at or near the skin temperature of their host, months after reproduction has begun) or H. lata though host physiology has been suggested to affect (peaking in winter when pigeons are not breeding) seasonal abundance of the blood-feeding ambly- are not coordinated with the onset or end of ceran louse, Menacanthus eurysternus (Burmeister) pigeon reproduction. The species on pigeons (Phthiraptera: Menoponidae) infesting European provide little support for the hypothesis that lice starlings (Sturnus vulgaris Linnaeus; Aves: Passer- populations are synchronised with the breeding iformes: Sturnidae). This species has been reported season of their host as is thought to be the case to cease reproduction during the winter months for a warbler (Foster 1969). Unlike the feral (Boyd 1951; Kettle 1983). This certainly was not the pigeon, this warbler is migratory with a short, case for the blood-feeding H. lata in our study, well-defined breeding season. which reached its greatest intensity during winter. Environmental factors may influence the Do seasonal changes in the life history of the seasonal patterns observed in our study. Moyer pigeon in Manitoba affect the seasonal dynamics et al. (2002a) showed that low relative humidity of the louse populations? Feral pigeons alternate had a detrimental effect on pigeon lice. In between roosting communally, when inter-bird Manitoba, there is little moisture in the air during contact is probably high, and breeding, when winter months, and this may cause the reduced contact is more restricted, primarily with a mate populations of C. columbae and C. compar we and offspring. These phases might affect dispersal observed at this time of the year. However, of lice to uninfested hosts, the onset of reproduc- relative humidity during the Manitoba winter is tion might change hormone levels in the bird that high, commensurate with low temperatures. It is cue louse reproduction (Foster 1969), or repro- not known whether such conditions affect louse duction might reduce the defence capabilities of populations. the pigeons (Hamstra and Badyaev 2009). In The seasonal moult cycle of pigeons might also Manitoba the seasonal biology of pigeons has not affect the seasonality of louse populations. Pigeons been investigated in detail. At least two broods per typically moult once per year (Johnston 1992), and year are reported, with eggs observed in nests in individual birds, the moult extends over a between 16 April and 29 July, and the latest prolonged period of many months (Johnston and nestling reported on 21 August (Taylor 2003). Janiga 1995). Since chewing lice are permanent However, our data for the occurrence of chicks ectoparasites that attach their eggs to feathers and show the nesting phase may begin in April some hair, moulting in the host may conceivably result in years, possibly earlier, and extends to October. some eggs being lost along with their substrate. Although sample sizes were small, chicks had Some authors have maintained louse populations lower prevalence of lice than juveniles, suggest- decline following the moult in their hosts (see ing that establishment of the lice on chicks Marshall 1981). However, Moyer et al. (2002b) does not occur immediately, perhaps not until demonstrated that experimentally induced moulting the development of feathers is adequate for in rock pigeons that had been bitted to prohibit the C. columbae (Johnston and Janiga 1995). impact of host preening had no detectable effect Campanulotes compar, however, had louse on populations of C. compar and C. columbae, intensities on chicks as high as those on juveniles, especially when the most efficient means of suggesting this species disperses to nestlings more estimating numbers of lice was used, i.e., washing. rapidly than the other louse species. By the time We washed pigeons in our study, so the numbers pigeons had fledged, the abundance of lice on the of lice recorded is a close approximation of the juvenile birds was similar to that on adults. numbers present on the birds.

Preening is perhaps the most important factor Wildlife Rehabilitation Centre for the care in affecting numbers of lice on any one pigeon. which they handled pigeons before reaching our Clayton (1991) has shown that when the effec- laboratory facility. Thanks also goes to Dave tiveness of preening is manipulated by placement Holder, Lisa Babey, and the many students during of a bit on the upper mandible of a pigeon, this study who helped by washing pigeons. This populations of C. columbae and C. compar research was funded in part by Discovery Grants to increased dramatically. Preening is affected by T.D.G. from the National Sciences and Engineering physiological and behavioural factors (Johnston Research Council of Canada. The continued and Janiga 1995) and no doubt by the presence of support of the Department of Entomology and organisms on the feathers and skin, so anything Faculty of Agricultural and Food Sciences, that results in changes in grooming behaviour University of Manitoba is gratefully acknowledged. might also impact numbers of lice. Despite this, we observed consistent seasonal patterns in louse infestations on feral pigeons. Preening by the host References may also affect competition between species of Bergstrand, J.L. and Klimstra, W.D. 1964. Ectopar- lice infesting pigeons. Bush and Malenka (2008) asites of the bobwhite quail in southern Illinois. The observed competitive displacement of wing lice American Midland Naturalist, 72: 490–498. by body lice, especially when preening ability of Boyd, E.M. 1951. A survey of parasitism of the starling Sturnus vulgaris L. in North America. Journal of the host was hampered. Galloway and Lamb Parasitology, 37:56–84. (2014) found no evidence for competition among Bush, S.E. and Malenka, J.R. 2008. Host defence chewing lice on pigeons, and it may be that louse mediates interspecific competition in ectoparasites. populations were not large enough in feral Journal of Animal Ecology, 77: 558–564. pigeons in our study for demonstratively compe- Chandra, S., Agarwal, G.P., and Saxena, A.K. 1988. Seasonal changes in the population of Mallophaga titive interactions to be expressed. on Acridotheres tristis. Angewandte Parasitologie, As is the case for their population stability 29: 244–249. (Galloway and Lamb 2014), the seasonal Clayton, D.H. 1991. Coevolution of avian grooming dynamics of these four lice are characteristic of and ectoparasite avoidance. In Bird-parasite interac- each species. The most similar species are the two tions. Edited by J. Loye and M. Zuk. Oxford University Press, London, United Kingdom. Pp. 258–289. most abundant ones, C. compar and C. columbae. Clayton, D.H. and Drown, D.M. 2001. Critical evalua- Their seasonal patterns of abundance differ, tion of five methods for quantifying chewing lice however, in the extent of the decline during winter (Insecta: Phthiraptera). Journal of Parasitology, 87: and the uniformity of their prevalence through the 1291–1300. year. Campanulotes compar is less seasonal than Eveleigh, E.S. and Threlfall, W. 1976. Population dynamics of lice (Mallophaga) on auks (Alcidae) C. columbae for all the parameters measured. The from Newfoundland. Canadian Journal of Zoology, two less common species, C. tovornikae and 54: 1694–1711. H. lata, have different seasonal patterns from the Foster, M.S. 1969. Synchronized life cycles in the two abundant ones. What causes these seasonal orange-crowned warbler and its mallophagan – changes in abundance remains to be determined. parasites. Ecology, 50: 315 323. Galloway, T.D. and Lamb, R.J. 2014. Abundance and Living on the surface of their warm-blooded host stability are species traits for four chewing lice allows these insects to be active all year round (Phthiraptera: Menoponidae, Philopteridae) on feral although the pigeons in this study were found in a pigeons, Columba livia (Aves: Columbiformes: north temperate environment with extreme annual Columbidae). The Canadian Entomologist, 145: – fluctuations in temperature. Nevertheless, three of 444 456. Galloway, T.D. and Palma, R.L. 2008. Serendipity with the four louse species show large and relatively chewing lice (Phthiraptera: Menoponidae, Philopter- consistent seasonal changes in abundance. idae) infesting rock pigeons and mourning doves (Aves: Columbiformes: Columbidae) in Manitoba, with new records for North America and Canada. The Canadian Entomologist, 140: 208–218. Acknowledgements Hamstra, T.L. and Badyaev, A.V. 2009. Comprehen- sive investigation of ectoparasite community and The authors thank the staff at the rehabilitation abundance across life history stages of avian host. hospitals of the Wildlife Haven and the Prairie Journal of Zoology, 278:91–99.

Johnston, R.F. 1992. Rock pigeon (Columba livia). The Rózsa, L., Reiczigel, J., and Majoros, G. 2000. Birds of North America, 13:1–14. Quantifying parasites in samples of hosts. Journal Johnston, R.F. and Janiga, M. 1995. Feral pigeons. of Parasitology, 86: 228–232. Oxford University Press, London, United Kingdom. Singh, S.K., Kumar, A., Badola, S.S., and Saxena, A.K. Kettle, P.R. 1983. The seasonal incidence of parasitism 2000. Site preference of four pigeon lice (Phthiraptera, by Phthiraptera on starlings (Sturnus vulgaris)in Insecta) on host body. Rivista di Parassitologia, 17: England. New Zealand Entomologist, 7: 403–408. 341–349. Marshall, A.G. 1981. The ecology of ectoparasitic Sokal, R.R. and Rohlf, F.J. 1981. Biometry. W.H. insects. Academic Press, New York, New York, Freeman and Company, New York, New York, United States of America. United States of America. Mironov, S.V. and Galloway, T.D. 2002. Four new SYSTAT 2010. SigmaPlot 11.2. SYSTAT Software, species of feather mites (Acari: Analgoidea). The Inc., San Jose, California, United States of America. Canadian Entomologist, 134: 605–618. Taylor, P. 2003. The birds of Manitoba. Manitoba Moyer, B.R., Drown, D.M., and Clayton, D.H. 2002a. Naturalists Society, Winnipeg, Manitoba, Canada. Low humidity reduces ectoparasite pressure: impli- Tuleshkov, K. 1965. Mallophaga (Insecta) – ectopar- cations for life history evolution. Oikos, 97: 223– asites on the starling (Sturnus vulgaris L.. Izvestiya 228. Zoologicheskiya Institut, 6: 175–187. (In Russian). Moyer, B.R., Gardiner, D.W., and Clayton, D.H. von Baum, H. 1968. Biologie und Ökologie der 2002b. Impact of feather molt on ectoparasites: Amselfederläuse. Angewandte Parasitologie, 9: looks can be deceiving. Oecologia, 131: 203–210. 129–175. Nelson, B.C. and Murray, M.D. 1971. The distribution Von Spitznagel, A. 1985. Unterschungen über den of Mallophaga on the domestic pigeon (Columba Befall der Wasseramsel (Cinclus c. aquaticus) met livia). International Journal for Parasitology, 1: Federlingen (Phthiraptera: Mallophaga). Ökologie 21–29. der Vögel, 7: 409–420. Petryszak, A., Rościszewska, M., Bonczar, Z., and Woodman, W.J. and Dicke, R.J. 1954. Population Pośpiech, N. 2000. Analyses of the population ﬂuctuations of the Mallophagan parasite Brüelia structures of Mallophaga infesting urban pigeons. vulgata (Kellogg) upon the sparrow. Wisconsin Wiadomości Parazytologiczne, 46:75–85. (In Academy of Sciences, Arts and Letters, 43: Polish). 133–135.