Composition and distribution of lice (Insecta: Phthiraptera) on Colombian and Peruvian birds: New data on louse-host association in the Neotropics

Abstract The diversity of permanent ectoparasites is likely underestimated due to the difficulty of collecting samples. Lice (Insecta: Phthiraptera) are permanent ectoparasites of birds and mammals; there are approximately 5,000 species described and many more undescribed, particularly in the Neotropics. We document the louse genera collected from birds sampled in Peru (2006–2007) and Colombia (2009–2016), from 22 localities across a variety of ecosystems, ranging from lowland tropical forest and Llanos to high elevation cloud forest. We identified 35 louse genera from a total of 210 bird species belonging to 37 avian families and 13 orders. These genera belong to two suborders and three families of lice: Amblycera, families Menoponidae (present on 131 bird species) and Ricinidae (39 bird species); and Ischnocera, family Philopteridae (119 bird species). We compared our bird-louse associations with data in Price et al. (2003) and recently published Neotropical studies. The majority of bird-louse associations (51.9%) were new, with most of these coming from Passeriformes, the most diverse avian order, with the most poorly known louse fauna. Finally, we found geographical variation in louse infestation and prevalence rates. With this study, we report the first comprehensive documentation of bird-louse associations for Colombia and substantially increase the known associations documented for Peru.


Introduction
Parasites are one of the most common forms of life on the planet (Price 1980). They have evolved repeatedly in every major clade (Poulin and Morand 2000). Although parasites are amongst the most diverse organisms in the world, few are well studied. Permanent ectoparasites are particularly difficult to study because they live their entire life cycle on hosts (Marshall 1981) and require capturing the host to sample them.
Lice (Insecta: Phthiraptera) are permanent parasites occurring on both birds and mammals. There are approximately 5,000 described species of lice, about 3,000 of which are known from birds (Price et al. 2003, Smith et al. 2011. The taxonomic diversity of lice is positively correlated with the taxonomic diversity of their hosts (Eichler 1942, Vas et al. 2012. Colombia and Peru harbor the richest avifaunas in the world (Jetz et al. 2012), with 1,878 and 1,852 bird species, respectively (Avendaño et al. 2017), and, correspondingly, the highest diversity of avian lice is thought to be found in these regions (e.g. Valim and Weckstein (2013)). Currently, however, there is limited knowledge of louse-host associations and louse diversity from these countries (e.g. Clayton et al. (1992) and the Neotropics in general (Clayton et al. (1992), Marini et al. (1996), Valim and Weckstein (2013). This is due in part to the poor representation of louse specimens in museum collections and the lack of louse specialists and field workers who sample parasites when collecting or handling birds. Therefore, the diversity of known louse species at regional scales is not on par with lists of avian host diversity from these countries. Our main objective is to provide novel information about the composition and distribution of lice on Colombian and Peruvian birds.
From large collections of louse specimens from birds in Peru, Clayton et al. (1992) and Clayton and Walther (2001) examined how host ecology and morphology influence louse diversity across a sample of 127 bird species. These two studies, amongst other taxonomic studies published using the same specimens e.g. Price and Clayton (1995), Price and Clayton (1989), provide most of the known louse-host associations from Peruvian birds. Much less information is available for Colombia, apart from the work of Melbourne A. Carriker (1879Carriker ( -1965, who collected mostly non-passerine birds and their associated lice and a study by Parra-Henao et al. (2011) where they identified lice from 18 bird species from the Cordillera Central near Medellín (Valle de Aburrá). Although this previous work provides an excellent starting point for understanding the diversity of lice in the Neotropics, the numbers of birds examined for lice is a small sample of the total avian diversity in this region.
In this study, we provide data from extensive sampling and description of louse-host associations from Colombia and Peru. Material was collected from 22 localities over nine years. From these samples, we identified 36 unique genera of lice and compared our results with those found in previous studies and with data compiled in the published checklist in Price et al. (2003). We found that over 50% of the louse-host associations were previously unreported and suggest that further data from these collections will be important to identify factors associated with louse diversity in the Neotropics. The data presented here provide the foundation for a long-term project sampling louse diversity across the Andes. This dataset will provide the basis for answering large-scale questions about patterns of diversity along elevational, habitat and host taxonomic gradients. The long-term project will include species level identification, taxonomic description and exploration of macro-ecological patterns along with archiving and storage of louse specimens.

Material and methods
Lice were collected at 22 localities in Peru (2006Peru ( -2007 and Colombia (2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016) (Table  1). In Peru, samples were collected by GAL and JEJ at four stations from Andean foothill forest (800 m a.s.l.) to high elevation cloud forest (3,000 m a.s.l.) inside Manu National Park or its buffer zone along a contiguously forested altitudinal gradient (Fig. 1a). In Colombia, samples were collected by GAL, JEA and JSP at 18 sites across the country, which ranged in elevation and habitat from 100 m a.s.l. to 2,800 m a.s.l., including savannah and gallery forest, lowland tropical forest and humid premontane and montane cloud forest (Fig. 1b) At each site, 10 to 20 netting stations were run and, at each station, 10 mist nets were opened for three days to capture birds. Each netting station was sampled twice during each 4 to 6 month field season. After removing birds from the nets, each individual host was placed in a clean cloth bag until processing for ectoparasites. We used three methods for collecting ectoparasites, detailed in Clayton and Drown (2001): 1) Post-mortem ruffling, 2) visual examiniation and, for the majority of samples 3) dust-ruffling. To dust-ruffle the birds, we applied ~1 ml of EverGreen pyrethrum dust (McLaughlin Gormley King Company, MN, USA) to captured birds and then ruffled feathers from all body regions except the head. Five minutes after we applied the powder, we ruffled each bird's feather tracts over a plastic sheet for 30 to 60 seconds to remove powder and ectoparasites. We transferred all powder a b and ectoparasites that fell on to the sheet to a 1.5ml Eppendorf tube filled with 96% ethyl alcohol and inserted a label with host metadata.
JEA also collected ectoparasite specimens using Clayton and Drown's (2001) post-mortem ruffling method for euthanised avian hosts. These hosts were collected and prepared as museum voucher specimens. To collect ectoparasites, JEA placed each euthanised host in a Ziploc bag with cotton soaked in ethyl acetate for 20 minutes. He then removed the bird from the bag and ruffled the plumage for 60 seconds over a white sheet of paper. Each specimen was returned to its Ziploc bag (with cotton soaked with ethyl acetate), ruffling the plumage two additional times, at intervals of 15 minutes. The ectoparasites were collected from the paper with a small brush and placed in a vial with 96% ethyl alcohol with a label including host specimen metadata. Bird voucher specimens were deposited in the bird collection of Instituto de Ciencias Naturales (ICN) of Universidad Nacional (Bogotá, Colombia) and the Museum of Natural History (MHNU) at Universidad de los Llanos (Villavicencio, Meta, Colombia). Lice were separated from the other ectoparasites, placed into individual vials and identified to genus using taxonomic keys Price et al. (2003). Host taxonomy followed the South American Classification Committee Remsen et al. (2017). Many louse species require microscopic examination of a slide-mounted specimen for species level identification. As this will be the focus of future work, these lice were only identified to genus. All specimens are stored at -80C for later DNA extraction and slide mounting at the Universidad Icesi, Cali Colombia. Vouchered, slide-mounted specimens will be made available at 1) Same as reported -avian species for which our study found the same louse genera as reported.
2) Fewer than reported -avian species for which our study found fewer louse genera than reported 3) More than reported -avian species for which our study found more louse genera than reported

Results
In Colombia, we sampled 1,032 individual birds from 280 species. Just over half, 51.6% (532), of these birds were infected with ectoparasites (i.e. feather mites, ticks, parasitic flies, fleas and lice) and we found lice on 30% (310) of individual hosts from 138 avian species, 36 avian families and 13 avian orders (Table 2). In Peru, we found lice on 262 individual birds from 98 species, 19 families and 5 orders (Table 3). In both countries combined, we identified 35 louse genera on 210 bird species from 37 avian families and 13 avian orders. Lice documented in this study are from two suborders and three families: Suborder Amblycera (Menoponidae and Ricinidae); and suborder Ischnocera (Philopteridae).

Columbiformes
Columbidae (   genera than previously reported. In addition, for 20 bird species (9.5% of the host species sampled), we found more louse genera than previously reported Fig. 2.

Data resources
The dataset is the result of several trips to 22 localities to study Neotropical bird communities in Colombia and Peru Table 1. In this study, we report lice on a total of 572 individual hosts totalling 210 bird species from 37 avian families and 13 orders. We identified 35 louse genera from two suborders and three families: Suborder Amblycera (Menoponidae and Ricinidae); and suborder Ischnocera (Philopteridae) Suppl. material 1

Discussion
In the present study, we report the genera of lice collected from 210 bird species at 22 sites in Colombia and Peru. We compared the louse-host association found in our study with the known genera of lice from these species of birds. We used Price et al. (2003), the most complete published bird-louse association list, along with recent Neotropical host-louse faunistic and taxonomic publications to assess the novelty of the host-parasite associations documented by our study. Bird-louse associations included in each category described in the methods above. The Y axis represents the number of bird species and the X axis indicates the categories in which bird species were grouped according to reported louse-bird associations.
We report 109 novel host-louse generic associations. This was not unexpected as we sampled several lowland and Andean habitats which have previously had few studies of bird-louse associations.
The majority (87.1%) of these new records were from Passeriformes. Knowledge of lice from many Passeriformes is relatively poor compared to non-passerines Sychra et al. (2007) and thus the diversity and number of undescribed parasites from these hosts is likely high e.g. Valim and Weckstein (2013) The distribution of lice is related to the distribution of their hosts Rózsa and Vas (2015) and many orders and families typically have parasites of distinctive louse faunas Smith (2001). Our data are consistent with generalised patterns across avian groups. For example, members of the Ricinidae are known to infect hummingbirds and small Passeriformes, whereas members of the Menoponidae are widely distributed across most avian families Rózsa and Vas (2015). Similarly, we found lice from the genus Ricinus on 36 species of Passeriformes from 11 host families. Myrsidea is a broadly distributed, mega-diverse genus Valim and Weckstein (2013), found mostly on Passeriformes and is considered to have a high degree of host-specificity Price and Dalgleish (2007). We also found that the louse genus Myrsidea, from the family Menoponidae, was distributed on 120 bird species, two of which were non-Passeriformes.
In Ischnocera, the family Philopteridae is widely distributed on birds Rózsa and Vas (2015). The various genera of Philopteridae are often specialised morphologically and behaviourally for living on a single microhabitat in the plumage (e.g. wing, head and/or body feathers) where lice can avoid host preening Johnson et al. (2012). This microhabitat specialisation may in part explain the host specificity and diversity of these lice. The three most common genera of Philopteridae found in our study were Philopterus, Brueelia and Rallicola. Of these, Philopterus was the most widely distributed genus in this family, occurring on a diverse array of passerine host families and a single non-passerine host species (42 bird species). Brueelia, the most speciose genus of lice in the family Philopteridae, infects avian hosts from many orders, including Coraciiformes, Passeriformes and Piciformes Weckstein (2011)Valim andWeckstein (2013) Gustafsson and Bush (2017). Similarly, we found Brueelia on 30 bird species, including two species of Coraciiformes, two species of Piciformes and 26 species of Passeriformes. Finally, the third most frequently collected genus was Rallicola, found on 18 bird species, including one host species in the order Charadriiformes and 17 host species in the order Passeriformes. Rallicola is one of the most speciose of ischnoceran louse genera and has been reported from the avian host orders Apterygiformes, Charadriiformes, Gruiformes and Passeriformes Price et al. (2003), Smith (2001).
Thirty percent of the Colombian birds sampled (138 host species) were infected with lice. In Peru, Clayton et al. (1992) found that 48% of birds examined (127 host species) were infected by lice, whereas in Brazil, Marini et al. (1996) and Oniki (1999) found that 20% of 313 individual birds (53 species) and 63% of 60 birds (38 species) had lice, respectively. Enout et al. (2012) found that 65% of 57 avian host species sampled were infected with lice. These studies suggest that louse prevalence may vary geographically. For example, for the flycatcher, Leptopogon amaurocephalus, in Brazil, Marini et al. (1996) and Oniki (1999), sampled two and one individual hosts respectively and all were infected with lice, whereas found two of three individuals sampled infected with lice. We found that in Remedios, Colombia, only 16.6% (n=12) of L. amaurocephalus individuals were infected. However, other host species had similar prevalence rates as reported in previous studies. For example, in Brazil, Oniki (1999) found that 67% of Turdus leucomelas sampled (n=3) were infected with lice and we found that all individuals of Turdus leucomelas sampled at two localities by us (n=4) were infected. However, a second Brazilian study conducted by Enout et al. (2012) found a 43% infestation rate (n=35) for the same bird species. It is difficult to determine the drivers behind variation in prevalence. It is possible that we are seeing an ecological pattern due to differences in humidity at the different sampling localities Moyer et al. (2002), Bush et al. (2009), host distributions or due to the different methods used by researchers to collect the lice. Additional work, examining sites where lice were collected with the same methodology, will help to address these issues.

Conclusions
This manuscript presents data on avian lice from 210 host species. We report and document significant new host-louse association records from poorly sampled yet diverse regions of the world. This information provides an important basis for future studies in the tropics and further enriches our knowledge of the parasite fauna associated with Neotropical birds.