The phylogeny of francolins (Francolinus, Dendroperdix, Peliperdix and Scleroptila) and spurfowls (Pternistis) based on chick plumage (Galliformes: Phasianidae)

What do we know about francolin and spurfowl chick plumage relationships?

Hall (1963) placed all spurfowls (occurring in Africa only) and francolins (occurring in Asia and Africa) under one genus, Francolinus, and described differences between chicks based on chick plumage. She divided the “monophyletic francolins” into eight taxonomic groups and mentioned that the chicks of the Bare-throated, Montane, Scaly, Vermiculated (all in the genus Pternistis), Striated (Dendroperdix) and Spotted (Francolinus) groups have a single conspicuous brown longitudinal band (with side lines) on the back through the crown (the crown cap) and a solid blackish stripe from the bill through the eye to the neck (extended eye stripe). The Red-winged (Scleroptila) group also have a variegated back, but with a narrower longitudinal band, and a narrower and darker crown patch than the spurfowls, with eight alternate black and white stripes on the sides of the crown, through and under the eyes.

Within the Red-tailed (Peliperdix) group, the female P. coqui chick is similar to the chicks of the Red-winged group in that it has a crown patch with lateral dark brown and buffy coloured stripes arranged across the back through the crown and stripes under the eyes. Conversely, the male P. coqui chick displays a similar pattern to the basic patterns displayed by the chicks of the Bare-throated, Montane, Scaly, Vermiculated, Striated and Spotted groups, namely a broad crown cap and a single conspicuous rufous longitudinal band across the back through the crown cap and a solid blackish eye line. This difference between male and female P. coqui constitutes sexual dimorphism at the chick stage (Hall 1963).

In the Red-winged group the striped facial pattern (disruptive plumage) becomes a permanent cryptic feature in the adults, but in spurfowls the facial eye stripes often disappear, resulting in a plain colour in many species (Little and Crowe 2011). Although Hall (1963) mentioned that some variation was evident among sub-species of F. francolinus and F. shelleyi, she did not elaborate on this. Overall, she concluded that small variations existed among species and groups.

Francolins and spurfowls are not monophyletic

Crowe et al. (1992, 2006) reviewed the monophyletic genus “Francolinus” proposed by Hall and found that it represents two distant clades split among five genera: spurfowls (Pternistis) and francolins (Francolinus, Scleroptila, Peliperdix and Dendroperdix). In other words, based on DNA and morpho-behavioural studies, francolins and spurfowls do not share a common evolutionary path (Milstein and Wolff 1987; Crowe and Little 2004; Crowe et al. 2006). Bloomer and Crowe (1998) included the genetic material of several species, and Mandiwana-Neudani et al. (2014) analysed the genetic material (in conjunction with vocalisations) of 40 species (18 francolins and 22 spurfowls) to demonstrate the dichotomy between spurfowls and francolins. Mandiwana-Neudani et al. (2011) described the syringeal features that separated the two groups. However, none of these authors assessed chick plumage in any depth.

We assessed the chick plumage characters of chicks of traditionally recognised spurfowl and francolin species, in conjunction with a broad selection of outgroup species, to (1) describe the chick plumage of spurfowls and francolins; (2) phylogenetically confirm or dismiss the dichotomy between spurfowls and francolins; (3) investigate the patterns of evolution of chick plumage characters which are diagnostic traced on a molecular phylogeny and (4) evaluate habitat matching of spurfowl and francolin chicks.

Phylogenetic analyses—chick plumage and molecular characters

Parsimony analysis was performed on both chick plumage and Cytb characters in Windows-based PAUP* ver. 4.0b10 (Swofford 2002) with the following tree search parameters considered: full heuristic search with all characters unordered and with equal weight, starting tree(s) obtained via stepwise addition; tree-bisection-reconnection branch-swapping option was in effect, and 1000 random additions of taxa (Maddison 1991). One tree was held at each step during stepwise addition, with branches collapsed (creating polytomies) if the maximum branch-length was zero. Multiple, equally parsimonious cladograms were recovered and strict consensus cladograms were constructed. Bootstrap resampling (Felsenstein 1985) with 1000 pseudoreplicates and five random additions of taxa per bootstrap pseudoreplicate was used to determine the degree of support for each node. Both the natal downy and Cytb phylogenies were rooted on M. maleo. Three measures of phylogenetic signal were gained by estimating consistency index (CI), which measures how well the data and the tree fit with each other, that is, the consistency of the tree to the dataset; the retention index (RI), which measures how well synapomorphies explain the tree; and the homoplasy index (HI), which measures the amount of homoplasy observed on a tree relative to the maximum amount of homoplasy that could theoretically exist. These metrics reveal the amount of phylogenetic signals inherent in both character matrices.

Ancestral state reconstruction

The ancestral state reconstructions were made on the existing mt Cytb phylogeny for eight natal downy characters (Table 3) using Parsimony approach implemented in Mesquite ver. 3.2 (Maddison and Maddison 2017). All character states were unordered and the analysis employed the Markov k-state 1 parameter (mk1) model (Lewis 2001) where any particular change from one state to the other is equally probable.

Broad groups of chick plumage

From the outgroup it was possible to select three broad groups of chick plumage colours and patterns by which to categorise the francolin and spurfowl chicks.

Plain: The darkish chicks of megapodes (Megapodiidae) are plain dorsally without stripes. The chick plumage of guineafowl species that live in the open understory of forests, such as Guttera pucherani and Agelastes meleagrides, are vaguely striped (rather blotchy) compared with the prominent striped patterns of guineafowl species that raise their chicks in grass, such as Numida meleagris and Acryllium vulturinum (Madge and McGowan 2002) (Fig. 2). The chick plumage of the forest-living A. niger is mainly dark rufous and black above (Little 2016).

Semi-striped: Grouse (Tetraonidae), Gallus gallus and Phasianus chicks have a variegated pattern with a central longitudinal band that proceeds through the crown to the forehead (Fig. 3). The rufous crown stripe often forms a single broad crown cap, with thinner trimmings on either side of the crown cap.

Striped: These chicks possess multiple stripes on the sides of their faces and are striped longitudinally across the crown right to the nares. Their back patterns are also variegated, but with a prominent central longitudinal band. Numida meleagris and A. vulturinum are typical examples of striped chicks (Fig. 2C, D). Coturnix coturnix chicks also display multiple stripes. The well-defined thin stripes across the crown and on the sides of the face of N. meleagris compared with the weakly defined stripes and blotches (fluidity) of A. meleagrides and G. pucherani suggest that the well-defined stripes emerged from a darkish plainer ancestor (see A. niger), with its black plain dorsal chick plumage. The latter three guineafowl species live in forests.

Chick plumage of spurfowl chicks

The 15 spurfowl chicks for which data were captured show a marked degree of similarity in terms of the selected character states (Table 3, Additional file 1: Table S1). The chick plumage of spurfowl chicks is regarded as semi-striped, since they lack the multiple disruptive stripes on the face and crown. The chick plumage of a hybrid P. adspersus and P. natalensis chick did not vary from this typical pattern. The back bears a broad, dark brown (rufous) central longitudinal band flanked by thinner and lighter-coloured stripes (otherwise known as a variegated back pattern). The crown has a broad rufous crown cap which is often trimmed by a thin, darker stripe on the edges. This crown cap is essentially an extension of the longitudinal band across the back. The face has an eye stripe which extends for a short distance in front of the eye (it often resembles a blotch in front of the eye). Some species, such as P. icterorynchus, P. nobilis and P. erckelii, generally have a reddish appearance, but their patterns (structurally) remain like the rest. P. hartlaubi is like the rest, with limited facial stripes and a variegated back pattern with a central longitudinal band on the back.

Chicks from two species showed prominent diagnostic features. Pternistis nobilis has a bold, thicker eye stripe, which is more prominent than the rest of the spurfowls, with a relatively thick frontal eye stripe. Pternistis erckelii has a solid, longish jaw stripe which in other species is faint or absent (Fig. 4).

Spurfowl chicks differ from their closest evolutionary relatives. Both the spurfowl chicks and Coturnix coturnix chicks are semi-striped on their heads but Alectoris chukar chicks possess stripes. Dorsally, only spurfowl chicks have a broad crown cap compared to scattered or banded crowns in the other two species (Table 3).

Chick plumage of francolin chicks

Data were captured for 11 species representing all four genera (Table 3).

Asiatic francolins: F. pondicerianus, F. francolinus and F. gularis are semi-striped. Like the spurfowls, they possess a rufous crown cap. The sides of their faces often have a few blotch-like blackish markings (probably remnants of stripes from an ancestor that was fully striped) behind the eye and on the neck. Like spurfowls, they also have a longitudinal band on the back with thinner, lighter-coloured lateral stripes (in other words, variegated back markings). Interestingly, only some F. pondicerianus chicks have a striped crown, showing intra-specific variability.

Dendroperdix sephaena is semi-striped, and shows resemblance to the spurfowls and Asiatic francolins. It has a rufous, crown cap with a solid eye line behind the eye, stretching down along the neck which is also typical for spurfowl chicks. It has a distinctive round dark spot on the neck, which it shares with F. francolinus and F. gularis (Fig. 5). The back also has a broad, dark brown central longitudinal band flanked with thinner, lighter-coloured stripes (in other words, variegated back markings) similar to spurfowl chicks.

The male P. coqui chick is semi-striped, with a rufous crown cap and a solid eye stripe behind the eye, a broad, dark brown longitudinal band on the back flanked by lighter lateral stripes reminiscent of spurfowl chicks (Fig. 6A, B). A female chick collected at Shangabue (Zimbabwe) shows a typical spurfowl crown and back, but the sides of the face have multiple stripes, which is typical of Scleroptila spp. (Fig. 6C). A similar female chick specimen is kept in the Ditsong National Museum of Natural History in South Africa. The distinction between female and male chicks is also evident at the juvenile stage, when the male has a clean buffy coloured face and the female multiple facial stripes; this is eventually carried through to adulthood. The male chicks resemble spurfowl chicks and are more closely related to the Asiatic chicks than to the striped chicks of Scleroptila spp.

Scleroptila gutturalis has a broad rufous crown cap, with multiple stripes on the face and a broad central longitudinal band on the back with thinner, light-coloured lateral stripes. Scleroptila levaillantii, S. shelleyi and S. afra have striped heads (instead of a well-defined crown cap) and facial markings, and the head colours are distinctively white and black, revealing a chevron pattern.

The Asiatic francolins, D. sephaena and the male P. coqui chick possess broad crown caps and semi-striped faces which are like their closest evolutionary relative G. gallus. Scleroptila species do not share these characteristics with G. gallus.

Chick plumage and Cytb phylogeny

The natal downy (Table 3) and Cytb data sets consisted of eight and 1,143 characters, consisting of eight and 430 parsimony informative characters, respectively. The matrices had a CI of 0.405, RI of 0.734 and HI of 0.595 whereas Cytb yielded a CI of 0.50, RI (of 0.521) and HI of 0.5, respectively. The Cytb phylogeny confirms the francolin-spurfowl dichotomy (Fig. 7) with francolins, spurfowls, grouse and allies (Tetraoninae) all in the family Phasianidae. Numididae is also recovered. The number of taxa included in this analysis had to match those for which the natal downy characters were available and this possibly could have resulted in the topology and nodal support that is slightly different from that in Crowe et al. (2006). The Asian/African linking lineage consisting of F. francolinus pondicerianus, F. gularis and D. sephana was recovered even though the branch from which these species arose is unresolved. The natal downy phylogeny (Fig. 8) shows poor phylogenetic structure which is largely polytomous and therefore no inferences can be made.

Symplesiomorphy underlying spurfowls

Ancestral characters form the “building blocks” that shape the colours and patterns of downy chicks. The broad longitudinal central rufous longitudinal band on the back and crown of spurfowl chicks (variegated) are also evident in G. gallus and grouse. The solid eye line of spurfowls, with the rest of the face often clean, is also evident in G. gallus, and is particularly similar between P. nobilis and G. gallus (see Figs. 3, 4). The prominent jaw stripe of P. erckelii (and to a lesser extent by other spurfowls) is also evident in Coturnix coturnix, N. meleagris and V. acryllium chicks.

Symplesiomorphy underlying francolins

The Asian francolins have a rufous central longitudinal band on the back and crown, like G. gallus. The blotches behind the eye (as in F. gularis) are like some Tetraonidae spp., being especially evident in D. obscurus pallides and L. lagopus. Dendroperdix sephaena has all the characteristics of the G. gallus chick. However, the single diagnostic isolated spot on the neck is similar to a spot recorded in the same area of the neck of F. gularis and F. francolinus (Fig. 5). Peliperdix coqui chicks also have the characteristics of G. gallus, but the female chick’s striped facial parts are similar to those of the N. meleagris chick. Scleroptila gutturalis falls in the same category as the female P. coqui chicks. It has a crown and back feather pattern like G. gallus, but the facial parts resemble N. meleagris chicks.

The chicks of S. afra, S. lavaillantii and S. shelleyi are similar to the striped pattern seen in N. meleagris, A. vulturinum and Coturnix coturnix chicks. However, the black lateral facial patterns of the latter three francolins stand out and could be related to the black facial parts of an Asiatic ancestor such as F. gularis. This is unlike S. gutturalis, which does not fall in this category and does not have black-coloured stripes. The plain colours of P. lathami chicks resemble the plain colours found in P. pucherani, A. meleagrides and A. niger.

Correlations between micro-habitat and chick plumage character states

Micro-habitat refers to the immediate environment of the bird within a broader landscape. Generally, species clades live in uniform micro-habitats for most of the year (Fig. 9). Some, such as N. meleagris and A. vulturinum live out in the open but breed in closed habitat, in other words in grass. Peliperdix lathami is an exception as it lives in forests but its close relatives, P. coqui and P. albogularis move into open micro-habitats (Fig. 9).

Most species that live in open micro-habitats (Peliperdix, Dendroperdix, Francolinus spp., Pternistis spp., L. lagopus, T. cupido and G. gallus) are semi-striped compared to Scleroptila spp., C. coturnix, N. meleagris and A. vulturinum that live in closed micro-habitats and are all multi-striped (more cryptic) (Additional file 1: Table S2). Most species that live in open micro-habitats possess rufous variegated colour combinations but species that live in closed micro-habitats have dark reddish and rufous colour combinations. Most forest species possess reddish colours (Additional file 1: Table S2).

Most species that live in open micro-habitats possess rufous caps on their crowns compared to all species living in closed micro-habitats that possess banded crowns. Forest species possess caps as well (Additional file 1: Table S2).

Most species that live in open micro-habitats possess a distinct eye band and comparatively most species that live in closed micro-habitats possess multiple facial stripes. Most forest species possess spots on the sides of their faces. Most species that live in open micro-habitats do not have jaw stripes whereas species that live in closed micro-habitats have partly or fully developed jaw stripes. Forest species do not possess any markings on the sides of their faces (Additional file 1: Table S2).

The back colours of species that live in open micro-habitats are rufous and most species that live in closed micro-habitats also possess rufous colours. Comparatively forest species possess dark reddish colours on their backs (Additional file 1: Table S2).

Most species that live in open micro-habitats possess broad stripes on their backs and species that live in closed micro-habitats possess thin and broad stripes. Comparatively forest species possess both but do not show any preference for any pattern (Additional file 1: Table S2).

Most species that live in open habitats possess no nose spots and for species that live in closed and forest habitats no preference was revealed. They both have spots or no spots. Generally, species of closed micro-habitats are more cryptic than species in open micro-habitats (Additional file 1: Table S2).

Ancestral state reconstruction

Phylogeny

From a cladistics point of view, the topology of the natal downy phylogeny of francolins and spurfowls is largely unsatisfactory or polytomous. This could be due to the high degree of symplesiomorphy within Phasianidae (Johnsgard 2008) (Fig. 8). It was therefore not possible to delimit species satisfactorily based on plumage characters. These results are incongruent with the spurfowland francolin phylogeny built on DNA and adult morpho-behavioural characters wherein resolved species relationships were determined with a high nodal support (Mandiwana-Neudani et al. 2014). The latter refutes Hall’s (1963) hypothesis and confirm that spurfowls (Pternistis) and francolins (Francolinus, Dendroperdix, Peliperdix and Scleroptila) do not form a monophyletic group (Crowe et al. 2006). Thus, the spurfowl and francolin phylogeny based on natal chick plumage characters does not support this dichotomy.

Inter-specific comparisons

Dendroperdix sephaena and F. gularis have a clear isolated dark spot on the side of the neck, a discontinued break between the eye stripe and the neck stripe, a faint jaw stripe and an eye liner which supports the view of Mandiwana-Neudani et al. (2014) that D. sephaena phylogenetically resembles the Asiatic francolins and might be considered a linking form between the Asiatic and African francolins (Fig. 5, Table 3).

Female P. coqui chicks have the disruptive eye stripes seen in S. gutturalis, and the male has limited facial markings as seen in the spurfowl chicks, which places this species between the two clades based on chick plumage. Hartlaub’s Spurfowl P. hartlaubi is regarded as the most basal species of all extant spurfowls, but has a mating system and vocalisations which differ substantially from all other spurfowls (Komen 1987; Van Niekerk 2013). There is some debate as to whether this species belongs to the genus Pternistis, but based on chick plumage it is similar to all spurfowls (J. Komen in litt). It was expected that the spurfowl chicks of P. ahantensis, P. squamatus and P. griseostriatus that live on the fringes of forests would have attained forest-like character states such as a dark, reddish, rufous, plain dorsal plumage or at the least vague blotches, instead of variegated stripes fusing into a darker background plumage colour. In other words, it was expected that the forest species would resemble other forest-living birds such as A. meleagrides, G. pucherani and P. lathami and the forest-dwelling tinamous chicks (Jehl 1971), all of which have plumage patterns which are not contrasting and generally darker. Instead they are very like all the other savannah-living spurfowls. Lack of forest-living characteristics among the spurfowls that live on the fringes of forests could be ascribed to a relatively recent speciation event, leaving these taxa in a state of flux, engaged in finding and refining habitat niches which may, in evolutionary time scales, imprint on chick plumage patterns and colours (Milstein and Wollf 1987). This lack of deviation from unity is supported by the polytomous structures in the cladograms (Fig. 8).

No character captured from spurfowl or francolin skins could be described as “derived” (not observed in any ancestor), which confirms the strong role that symplesiomorphy plays in natal chick plumage within Phasianidae, as was demonstrated within grouse (Short 1967) and which also seems applicable to other families such as that of the tinamous (Bertelli et al. 2002). In fact, at the outset it was demonstrated that spurfowls and francolins could each be divided into three broad plumage categories found in the wider Phasianidae family, namely plain, striped and semi-striped chicks, which confirms the role played by symplesiomorphy.

Field identification

It was possible to separate subsets according to plumage characteristics. Scleroptila show strong signs of disruptive lateral facial patterns. Conversely, the spurfowls have limited facial stripes, but a broader longitudinal band running from the crown down the neck and through the centre of the back. These differences have also been reported in the literature (Little 2016). The Asian francolins have a broad longitudinal band on the back like the spurfowls, but have black spots or blotches (speckled remains of stripes) just behind or below the eye, which is more in accordance with the spurfowls than for Scleroptila spp. Asian francolins are similar to D. sephaena, which has a single round dark spot just below the ear.

Pternistis erckelii is similar to the rest of the spurfowls, but has a distinct jaw line which is only partly visible on the jaws of a few other spurfowls. The plain back plumage of Peliperdix lathami (formerly F. lathami; Hall 1963; Madge and McGowan 2002) is at variance with the spurfowls and francolins that have dorsal stripes. These plain dorsal features on the back could be an adaptation for life in forest conditions to break the bird’s outline in shade, whereas semi-striped or striped plumage might be detected by predators under these conditions (Stevens and Merilaita 2008). A more complete set of francolin and spurfowl skins than what is currently available may reveal useful diagnostic features for individual species, but it is doubtful whether it would reveal derived traits that would assist in describing speciation based on chick plumage.

Habitat matching

Generally, species that live in open micro-habitats possess less cryptic markings than species that live in closed micro-habitats or forests. Coturnix coturnix is closely related to Pternistis spp. but unlike Pternistis spp. they live in grasslands and therefore possess cryptic multiple stripes over their faces and crowns like the chicks of Scleroptila spp. to which they are not as closely related phylogenetically (Fig. 7).

Scleroptila species. are not closely related to Numida meleagris or Acryllium vulturinum but as the latter species raise their chicks in grass like Scleroptila spp., they all possess cryptic stripes like Scleroptila spp. Peliperdix lathami lives in forests and therefore does not possess an eye stripe like Peliperdix coqui, Peliperdix algularis, Francolinus spp. and Pternistis spp. that live in open micro-habitats. Peliperdix spp., Dendroperdix sephaena and Francolinus spp. are more closely related to Scleroptila spp. but unlike Scleroptila spp. they live in open macro-habitats and therefore unlike Scleroptila spp., they do not possess multiple stripes across their faces and crowns, but instead possess a single wide eye stripe like Pternistis spp. which live in open micro-habitats. Gallus gallus are more closely related to Scleroptila spp. but their chicks look like the chicks of Pternistis spp. since they also live in open micro-habitats. It is therefore concluded that habitat matching is a likely driving force that determines the natal down of francolin and spurfowl chicks.

Does the chick plumage of spurfowl and francolin chicks have any significance for camouflage? Why don’t these chicks simply have plain colours like A. meleagrides or Peliperdix lathami that live under the canopy of forests? The limited facial stripes of spurfowl chicks probably assist the chicks to blend in with a bushier type habitat with limited ground cover such as grass. Should these chicks possess multiple stripes it would be conspicuous under these open conditions as it would define the outline of the head (Merilaita and Lind 2005; Stevens and Merilaita 2008). On the other hand, the multiple facial stripes of S. afra, S. levaillantii and S. levaillantoides chicks possibly break the outline of the shape of the chick’s head in grass cover. A non-striped head will be more detectable in grass than is a striped head (Merilaita and Lind 2005). In both francolins and spurfowls, the variegated back stripes probably provide countershading, making it difficult for a predator to distinguish the shape of the chick from its physical surroundings (Caro 2014).

In summary, although symplesiomorphy is a hereditary explanation for downy colours and patterns, the traits relevant for habitat matching are combined in a manner which is determined (adaptation) by natural selection (Frost 1975; Paxton 2009; Peichel 2014).

Evolution of downy patterns

Macrocephalon maleo is the most basal outgroup species to spurfowl and francolin within the Galliformes. Its downy chick plumage is basically uniformly black-brownish. The divergence of other downstream variants is basically a dilution of darker pigment (Mills and Patterson 2009). While speciation continued dark plain colours became limited but patterned on certain body parts of descendants, particularly on the crown, face and back. A. meleagrides and Guterra pucherani are proximate descendants to M. maleo and here it is evident that the uniformly dark plumage of downy chicks possesses whitish (diluted) blotches, but not yet particularly sharply defined features (e.g. striped).

Further down the evolutionary pathway two overriding morphological forms emerged with less black-brownish pigmentation in certain areas revealing dark stripes on their heads and back that are symmetrical and orderly arranged on a buff-coloured (diluted) background. These patterns are observed among many Phasianidae but particularly among Pternistis spp. where it emerged as semi-stripes (fewer stripes) compared to Scleroptila spp. where the remaining rufous stripes emerged as multiple stripes, giving the bird a disruptive chick plumage.

The presence of nose spots happened when most darker pigmentation receded leaving two isolated spots (or short stripes) on either side of the base of the upper mandible (nares). Indeed, in some species such as Lagopus lagopus, Scleroptila afra and Numida meleagris there are continuous stripes from the nose across the orbit on both sides of the head showing that dilution emerged along stripes. The presence of jaw stripes and multiple stripes on the face is ascribed to the evolution of darker pigmentation in this area. An ancestral bird like M. maleo has uninterrupted dark plain colours all over the head but the darker pigmentation diluted on the lower facial areas, but leaving no distinguishable jaw stripe (Additional file 1: Fig. S1).

There are more stripes on the side of the face of Pternistis erckelii than in other spurfowls, such as having a distinct jaw stripe and a thin stripe above the orbit which is more bold and typical of Scleroptila spp. and female P. coqui. The jaw stripe is only faintly visible on some other spurfowls. Ostensibly, while this seems to be a logical course of evolutionary events, for instance that P. erckelii could link the spurfowls and Scleroptila spp., it is more likely that the ancestors of extant francolins and spurfowls moved back and forth between striped and semi-striped chick plumage, which was probably determined by exposure to different habitats.