Two new species of socially parasitic Nylanderia ants from the southeastern United States

Abstract In ants, social parasitism is an umbrella term describing a variety of life-history strategies, where a parasitic species depends entirely on a free-living species, for part of or its entire life-cycle, for either colony founding, survival, and/or reproduction. The highly specialized inquiline social parasites are fully dependent on their hosts for their entire lifecycles. Most inquiline species are tolerant of the host queen in the parasitized colony, forgo producing a worker caste, and invest solely in the production of sexual offspring. In general, inquilines are rare, and their geographic distribution is limited, making it difficult to study them. Inquiline populations appear to be small, cryptic, and they are perhaps ephemeral. Thus, information about their natural history is often fragmentary or non-existent but is necessary for understanding the socially parasitic life history syndrome in more detail. Here, we describe two new species of inquiline social parasites, Nylanderia deyrupisp. nov. and Nylanderia parasiticasp. nov., from the southeastern United States, parasitizing Nylanderia wojciki and Nylanderia faisonensis, respectively. The formicine genus Nylanderia is large and globally distributed, but until the recent description of Nylanderia deceptrix, social parasites were unknown from this genus. In addition to describing the new social parasite species, we summarize the fragmentary information known about their biology, present a key to both the queens and the males of the Nylanderia social parasites, and discuss the morphology of the social parasites in the context of the inquiline syndrome.


Introduction
Ant social parasites exploit the social colony structure of free-living ant species, and they rely on their hosts for colony founding, survival, and reproduction for at least a part, and frequently the entirety of their life-cycles (Hölldobler and Wilson 1990;Buschinger 2009). Social parasitism is a life history strategy exhibited by at least 300 species among the approximately 13,500 described extant ant species. Traditionally, three main types of social parasitism have been recognized: temporary, dulotic, and inquiline social parasitism (Wasmann 1891;Wheeler 1910;Bourke and Franks 1991;Buschinger 2009). Inquilinism has evolved independently many times in the ants, and approximately 100 species are known from at least 30 ant genera which are distributed across six different subfamilies of the formicoid clade. Most inquilines do not produce a worker caste, and instead they invest their reproductive effort in producing sexual offspring. Many inquilines have convergently evolved a suite of similar morphological characteristics known as the "inquiline syndrome" (sensu Kutter 1968;Wilson 1971). These shared characteristics often include: elongated antennal scapes, reduced mouthparts, reduced body size, smooth and shiny cuticle, reduction or absence of the worker caste, intranidal mating with close relatives (i.e., adelphogamy), and polygyny (Kutter 1968;Wilson 1971Wilson , 1984Hölldobler and Wilson 1990;Radchenko and Elmes 2003;Rabeling and Bacci 2010;Rabeling et al. 2015Rabeling et al. , 2019Bharti et al. 2016). Interestingly, independently evolved inquiline species exhibit a mosaic of inquiline syndrome characteristics, frequently converging on a similar albeit not identical parasitic phenotype (Wilson 1984;Radchenko and Elmes 2003;Rabeling and Bacci 2010;Rabeling et al. 2019).
The genus Nylanderia is a member of the ant tribe Lasiini in the subfamily Formicinae (Blaimer et al. 2015) and presently consists of 150 described taxa (Bolton 2019;LaPolla and Kallal 2019). The genus is globally distributed, with the majority of species being found in warm, forested regions, though it is largely absent from the temperate regions of the Palearctic (LaPolla et al. 2011;Bolton 2019;LaPolla and Kallal 2019). In the Nearctic region, 14 native and six introduced species are recognized (LaPolla et al. 2010;Kallal and LaPolla 2012), representing a rather modest fauna given the high diversity and large biogeographic extent of the genus. Until recently, social parasitism was unknown in Nylanderia ants, and the first inquiline social parasite in the genus, Nylanderia deceptrix (Messer et al. 2016), was described from Massachusetts.
Here, we describe two new Nylanderia inquiline social parasites from the Nearctic and provide keys for identifying them. In addition, we summarize our current knowledge about the biology and natural history of these social parasites, and we briefly discuss the species morphologies and life histories in the context of the inquiline syndrome as well as inquiline evolutionary biology.

Material examined ABS
Archbold Biological Station, Venus, FL, USA; MCZC Museum of Comparative Zoology Collections, Harvard University, Cambridge, MA, USA; SIBR Social Insect Biodiversity Repository, Arizona State University, Tempe, AZ, USA; USNM National Museum of Natural History, Washington, DC, USA.

Morphological analysis
Specimens were measured at the MCZ using a Wild M5A stereomicroscope (100× magnification) fitted with an ocular micrometer. Measurements were recorded and rounded to the nearest 0.01 mm at the highest magnification possible for each measurement and specimen. Composite images were generated at ASU using a Leica DFC450 digital camera mounted to a Leica M205 C stereomicroscope and assembled using Leica Application Suite (Version 4.5) and Helicon Focus (Version 6.6.1) software packages. Measurement terminology, abbreviations, and definitions follow LaPolla et al. (2011) and Kallal and LaPolla (2012):

Statistical analysis of morphological measurements
To quantify morphological differences characteristic of the inquiline syndrome in Nylanderia ants, we collected morphological measurements for social parasites and their hosts and analyzed them statistically. We measured Weber's Length (WL) as a proxy for Total Length (TL) because the gaster of individuals was often damaged during collection. Statistical analyses were conducted using R 3.4.0 (R Core Team 2017) statistical package. Due to low sample sizes that likely contributed to the data being non-normally distributed, we used Kruskal-Wallis tests to determine whether hosts and parasites were significantly different in size. In addition, we applied pairwise Mann-Whitney tests post-hoc to determine individual differences between social parasites and their respective hosts, as well as between the social parasites. A Bonferroni correction was applied to the Mann-Whitney tests to account for multiple comparisons between species and castes, and to provide a more conservative alpha to compensate for low sample sizes in some cases (see below). For the three Nylanderia host parasite pairs, we analyzed the following samples: N. deyrupi: 29 queens and 5 males; N. parasitica: 7 queens and 10 males; N. deceptrix: 22 queens and 5 males; N. parvula (Mayr, 1870): 19 queens, 13 males, and 15 workers; N. wojciki (Trager, 1984): 17 queens, 8 males, and 20 workers; and N. faisonensis (Forel, 1922): 17 queens, 10 males, and 29 workers. Morphological measures of N. deceptrix were taken during an earlier study (Messer et al. 2016), and we added the morphological measurements of free-living species reported by Kallal and LaPolla (2010) to our dataset.
Etymology. This species is named in honor of Mark Deyrup, who first discovered the miniature females of N. deyrupi in malaise trap samples. Mark Deyrup has been a resident naturalist at Archbold Biological Research Station in Central Florida since 1982, and he is a uniquely gifted natural historian who acquired a phenomenal knowledge about the biology of the ants of Florida. Mark recently synthesized his knowledge in the richly illustrated monograph on the Ants of Florida (Deyrup 2016). His meticulous studies of ant natural history and taxonomy have inspired students and colleagues alike for decades, and without Mark's insightful studies the rich natural history of Florida would be much less explored.
Distribution and natural history. Nylanderia deyrupi is a rare, apparently workerless inquiline social parasite occurring only in nests of its host, Nylanderia wojciki. It is similar in morphology, and apparently in life-history, to Nylanderia deceptrix, the inquiline parasite of N. parvula. Its host, N. wojciki is native to Florida and the adjacent southeastern states. It is a common ant in sandhill and pine flatwood communities. In contrast, N. deyrupi is presently known only from Archbold Biological Station and two areas east of Sebring in Highlands County, Florida ( Fig. 9; see also Deyrup 2016), all of which are located on the Lake Wales Ridge in central Florida. This sand ridge is more than one million years old (Turner et al. 2006), and it is home to endemic plants and animals, all narrowly distributed on the ridge itself. Deyrup (2016) suggested that N. deyrupi (referred to as Nylanderia Species A in Deyrup 2016) may be another such endemic species.
The host, N. wojciki, makes small (< 300 workers), usually monogynous colonies nesting in leaf litter or sand, usually in partly or lightly shaded areas (Trager 1984). The nests are often diffuse in the summer months, consisting of multiple shallow chambers within an area of 1-2 square meters, containing workers, brood, and sometimes sexuals. Collections of N. deyrupi consisted of queens and males scattered among these small host nest pockets, a pattern extremely similar to that seen in N. deceptrix and its host N. parvula. A striking feature of its life history, also shared with N. deceptrix, is that, unlike many ant inquilines, N. deyrupi does not appear to suppress the development of host sexuals. In collections made by M. Deyrup and S. Cover both host and parasite sexuals were commonly found together in the host nest, along with host worker brood and callows, strongly suggesting that the host queen is retained, not eliminated in parasitized colonies, coexisting with the social parasite. Another interesting life-history trait shared with N. deceptrix is the production of apterous males, which is unique among Nylanderia ants. Accordingly, males have limited mobility and probably no dispersal capability, and mating is expected to occur in or around the host nest (i.e., adelphogamy). Considering the limited mobility and the small number of males present in each nest, inbreeding is expected to occur in N. deyrupi. In addition, N. deyrupi has a strongly female-biased sex ratio, a phenomenon that has been frequently observed among inquiline social parasites.
Etymology. Nylanderia parasitica inhabits the nests of N. faisonensis, exhibits morphological characteristics of the inquiline syndrome, and potentially lacks a worker caste. Hence, the species epithet is indicative of the socially parasitic life history of N. parasitica.
Distribution and natural history. Information on the natural history and biogeography of N. parasitica is extremely limited. In previous publications, N. parasitica was referred to as undescribed socially parasitic Nylanderia species (N. sp. 1) (Kallal and LaPolla 2012) and Nylanderia Species B (Deyrup 2016). Most individuals of N. parasitica were collected from Hamilton County, Florida. Two alate queens were collected from Alachua County, Florida, and a single alate queen was collected from Jones County, Georgia ( Fig. 10; see also Deyrup 2016). The type series was collected inside the nest of N. faisonensis in a rotten log located in an upland oak-pine hammock and a pond swamp area in Hamilton County (Deyrup 2016). Unfortunately, no additional observations were recorded from this mixed colony. The two queens from Alachua County were collected in malaise traps in May and September, suggesting that N. parasitica females disperse on the wing throughout the warm summer months.
The host of N. parasitica, N. faisonensis, is widely distributed in the southeastern United States (Kallal and LaPolla 2012) and is primarily a woodland species living in rotting branches, under rotting tree bark, or in the leaf litter (Trager 1984). Nests are often fragmented, and they do not make a soil nest like most other Nearctic Nylanderia species, with the single queen living deep under the leaf litter, while workers and brood live near the surface (Trager 1984). Alates are reared from August to December, followed by an overwintering period typical of Nearctic Nylanderia. Alate individuals disperse from the maternal nests between March and May, while more southern populations start dispersing earlier during those months (Trager 1984). It remains unknown whether N. parasitica is tolerant of the N. faisonensis queen.
Worker caste. Our limited collections of N. parasitica have not yielded any putative workers for this species. Therefore, like N. deceptrix and N. parasitica, it is likely that this species is a workerless inquiline.
Comparing the social parasite males to the males of their respective host species, N. deceptrix and N. deyrupi were not significantly different in body size from the host males (Mann-Whitney tests: P = 0.44 and P = 1, respectively). In contrast, N. parasitica males were significantly smaller than N. faisonensis males (Mann-Whitney test: P = 0.01). When males of the three social parasite species were compared to each other, no significant difference in size was detected (Kruskal-Wallis test: χ2 = 2.67, df = 2, P = 0.26).
Wing size. Behavioral observations revealed that queens and males of N. deceptrix do not mate or disperse on the wing (Messer et al. 2016). Both inside nest mating and a reduced dispersal ability are important life history traits of inquiline social parasites, contributing to their localized distribution and frequently inbred population structure. Therefore, and as a proxy for flight performance, we measured the wings lengths of 13 free-living and three socially parasitic Nylanderia species. To test whether the social  parasites have shorter relative wing lengths when compared to free-living Nylanderia species, we calculated the ratio of Forewing Length to Weber's Length and compared the values across Nearctic Nylanderia species for both queens and males. Significant differences between species were detected in both queens (Kruskal-Wallis test: χ2 = 140.46, df = 15, P < 2.2 × 10 -16 ) and males (Kruskal-Wallis test: χ2 = 71.748, df = 13, P = 3.819 × 10 -10 ). Pairwise Mann-Whitney tests determined that the wing sizes of N. deceptrix and N. deyrupi queens were significantly reduced relative to their host species (P = 2 × 10 -7 ; P = 2.2 × 10 -4 , respectively; Fig. 8A). In contrast, both queens and males of N. parasitica did not exhibit any significant reduction in wing size relative to the host N. faisonensis (Mann-Whitney tests: P = 1; P = 0.364, respectively; Fig.  8B). Males of N. deceptrix and N. deyrupi were not included in the pairwise analysis, because they are brachypterous and apterous, respectively (Figs 1, 3).

Discussion
We described two new workerless inquiline social parasite species in the genus Nylanderia, N. deyrupi and N. parasitica, from the southeastern United States. Nylanderia deyrupi was discovered in nests of N. wojciki, and N. parasitica was found once inside the nest of N. faisonensis. In ants, the presence of mixed colonies is indicative of a socially parasitic life history. Nylanderia deyrupi was collected repeatedly at or around Archbold Biological Station in central Florida, which yielded first insights into the biology of the species. In contrast, very little information is known about N. parasitica, which was only observed alive once in northern Florida. Therefore, our interpretations regarding the biology of the two species, especially of N. parasitica, should be regarded as preliminary and would greatly benefit from additional study. Notwithstanding, first observations suggest that N. deyrupi is a workerless, host queen tolerant inquiline because the N. wojciki queen, callow workers, and sexual brood were found inside the host colonies, whereas workers of N. deyrupi were absent. Nylanderia parasitica was only found in a mixed colony with N. faisonensis, and at the moment it remains unknown whether this inquiline species is host tolerant or not, but workers of N. parasitica were also absent from this mixed colony.
The description of these two social parasite species increases the diversity of Nearctic Nylanderia to 17 species, and three of them are inquiline social parasites. Approximately 100 species of inquiline social parasites are known from six ant subfamilies. The majority of the inquiline social parasites belong to the subfamily Myrmicinae, and only 12 inquiline species are known from the subfamily Formicinae, including the genera Anoplolepis, Camponotus, Cataglyphis, Formica, Nylanderia, Plagiolepis, and Polyrhachis (Hölldobler and Wilson 1990;Buschinger 2009;Karaman 2012;Casevitz-Weulersse 2014;Messer et al. 2016). Considering that inquiline social parasites are less common in formicine ants, these new Nylanderia inquiline species provide an opportunity for comparatively studying the morphological, behavioral, and ecological traits associated with inquiline social parasite evolution in formicine ants.
Nylanderia deyrupi and N. parasitica seem to have limited geographic distribution ranges, which is typical for inquiline species (Wilson 1971;Buschinger 2009). So far, N. deyrupi is known from central Florida (Fig. 9), while N. parasitica was collected in northern Florida and southern Georgia (Fig. 10) (Deyrup 2016). These distribution ranges are significantly smaller than the ranges of their respective host species (Figs 9, 10) (Trager 1984;Kallal and LaPolla 2012;Deyrup 2016). However, N. parasitica has a considerably larger known range compared to both N. deceptrix and N. deyrupi. Considering that males of N. parasitica are fully winged and that queens were collected in Malaise traps, it is possible that mating and/or dispersal flights occur in this species, which could contribute to a wider geographic distribution. The currently recognized biogeographic distribution almost certainly also reflects sampling biases, considering that inquiline social parasites are rarely found. Figure 9. Geographic distribution of N. deyrupi (black star) and its host N. wojciki (red circles). Host distribution data was supplemented with additional information from antmaps.org (Janicki et al. 2016).

The inquiline syndrome of Nylanderia social parasites
Social parasites display a mosaic of morphological, behavioral, and life history traits characteristic of their socially parasitic biology, known as the inquiline syndrome (Kutter 1968;Wilson 1971). Nylanderia inquiline social parasites show adaptations and losses associated with a socially parasitic life history, including a loss of the worker caste, polygyny, elongated scapes, lighter coloration, reduced body sizes, reduced wings, and a loss of antennal segments (Table 1). Other inquiline syndrome characters outlined by Wilson (1971) and Hölldobler and Wilson (1990), such as reduced labial and/or maxillary palps, a smooth and shiny cuticle, and a reduced pilosity could not be observed in Nylanderia social parasites, supporting the hypothesis that morphological, behavioral, and life history traits characteristic of a socially parasitic lifestyle evolve convergently in a mosaic fashion (Wilson 1984;Hölldobler and Wilson 1990;Radchenko and Elmes 2003;Rabeling and Bacci 2010;Rabeling et al. 2015Rabeling et al. , 2019. We briefly discuss the most significant modifications observed in Nylanderia inquiline social parasites. Body size reduction. In comparison to their hosts, all three Nylanderia social parasite species are significantly reduced in size. A comparative analysis of the inquiline syndrome in Pheidole and fungus-growing ant social parasites revealed that body size reduction is one of the first traits to evolve in inquilines (Wilson 1984;Rabeling and Bacci 2010). Nonacs and Tobin (1992) conducted an analysis of inquiline size relative to their hosts, using head size as a proxy for body size, and discovered that the queens of 18 of the 19 species in their study were equal in size or smaller than the host workers. A behavioral study of Plagiolepis inquilines examined the effect of size reduction on social parasite survival, revealing that miniaturization prevented P. xene queen and male brood from being culled by host workers (Aron et al. 1999(Aron et al. , 2004. In contrast, P. pygmaea host males, which are larger than the host workers, were actively removed Figure 10. Geographic distribution of N. parasitica (black stars) and its host N. faisonensis (red circles). Host distribution data was supplemented with additional information from antmaps.org (Janicki et al. 2016). Table 1. Comparison of inquiline syndrome characteristics for N. deceptrix, N. deyrupi, and N. parasitica. Traits applying to females but not males are marked with an asterisk (*), whereas traits applying to males but not females are marked with a cross ( + ). Morphological reductions observed in social parasites were determined by comparisons relative to the respective host species. cies, however, and when sufficient samples become available, future studies need to test directly for population genetic signatures of inbreeding in Nylanderia social parasites.
Reduction of antennal segments. One trait of the inquiline syndrome that is unique to N. parasitica and absent from N. deceptrix and N. deyrupi is the reduction in the number of antennal segments from 13 to 12 in males. A reduction of antennal segments has been observed in some social parasite species of fungus-growing ants, such as Pseudoatta argentina and Mycocepurus castrator (Gallardo 1916;Rabeling and Bacci 2010). The reduction of antennal segments is potentially correlated with a reduced number of olfactory receptors, but this hypothesis remains to be tested.

Outlook
With currently three known social parasite species, the genus Nylanderia developed into an interesting study system for exploring the evolutionary history of social parasitism in a comparative context. In general, inquiline social parasites are of interest to evolutionary biology because of their departures from a free-living life history, the convergent morphological and behavioral evolution of traits associated with the socially parasitic life history, as well as their close phylogenetic relationships to their hosts. Previous studies revealed that some inquiline species evolved directly from their host species via sympatric speciation (Savolainen and Vepsäläinen 2003;Rabeling et al. 2014;Leppänen et al. 2015;Nettel-Hernanz et al. 2015) whereas other inquilines likely originated in allopatry (Agosti 1994;Sanetra and Buschinger 2000;Ward et al. 2015). In a forthcoming study, we will test whether Nylanderia inquilines evolved via the intra-or the interspecific route of social parasite evolution. In addition to inferring the evolutionary history, it is critical to study the behavior and natural history of Nylanderia inquilines to gain a more detailed understanding of their biology.