Genetic Identity of the Red-legged Partridge (Alectoris rufa, Phasianidae) from the Island of Madeira

Simple Summary

We wanted to determine to which subspecies the Red-legged Partridge living on Madeira, Portugal belongs to, and which Iberian populations are closest to it. This medium-sized galliform, which is endemic to southwestern Europe, was introduced into this island for hunting purposes in the XV century. We sequenced the mitochondrial DNA of Madeiran partridges, relying on either fecal samples collected in the wild or specimens dating from 1900 to 1964 and preserved in museum collections. Then, we compared our results with an already published sequence dataset covering the entire Iberian range of the species. We proved that the population from Madeira belongs to the Galician Partridge, with northern Portugal and northwestern Spain being the most likely sources of its historical introduction into the island. We found no significant change in the genetic pattern of Madeiran partridges over time as well as no evidence for the occurrence of genes from other congeneric species such as, for instance, the Chukar Partridge. This characterization provided an important piece of knowledge for the management of a popular hunting resource of Madeira.

Abstract

The Red-legged Partridge (Alectoris rufa, Phasianidae) is a non-migrant gamebird endemic to southwestern Europe that was introduced into Mediterranean and Atlantic islands in historical times. This is the case for Madeira, Portugal, where a population morphologically assigned to A. r. hispanica has been present since the XV century. We assessed its genetic identity using 2248 (Cytochrome-b, Cyt-b + Control Region, CR) and 297 bp-long (CR) mitochondrial DNA sequences obtained from modern and archival (1900–1964, including Caccabis rufa maderensis syntypes) partridges, respectively. These sequences were compared against an already published dataset covering the entire Iberian A. rufa range. We found that all the haplotypes of modern birds from Madeira were private to this island. The putative subspecies was confirmed, and northern Portugal with northwestern Spain turned out to host the closest mainland populations. This result was in line with the origin of the first human settlers of Madeira from, among other historical provinces, Douro Litoral and Minho, the latter neighboring Galicia. Despite relatively recent A. rufa importations from continental Europe, we did not find any significant change over time in the haplotypic pattern of Madeiran partridges as well as any evidence for maternal introgression from species such as the congeneric Chukar Partridge (A. chukar). Studies relying on genome-wide markers and including the only captive-bred population of Madeira are needed to gain more comprehensive information for the management of the local A. rufa.

1. Introduction

Humans have moved wild species beyond their native ranges for millennia [1], with birds historically playing an important part in this faunal dispersal that has significantly affected the present-day distribution of biodiversity. The Alectoris partridges (Phasianidae), which are medium-sized gamebird ranging across the entire Palearctic, have been heavily managed by man and intentionally translocated since thousands of years [2], as they represented important resources for hunting, farming (eggs, chicks), and medicine, serving also as pets and diplomatic gifts other than regularly featuring in ancient art [3]. Being a sedentary species with prevailing ground-dwelling behavior, the populations living on several Mediterranean and northern Atlantic islands are, indeed, the result of human-mediated introductions carried out for hunting purposes in historical times [4]. For instance, the Chukar Partridge (A. chukar) was translocated not later than 4000 years ago [5,6] to Cyprus and the Aegean islands, while the Barbary Partridge (A. barbara) was moved from North Africa to Sardinia (Italy) at least 2000 years ago [7] and to Tenerife (Canary Islands) in the XV century [8].

The Red-legged Partridge, Alectoris rufa, includes three subspecies established on the basis of subtle morphological differences in the color pattern [9,10]: A. r. hispanica (Seoane, 1894), the so-called Galician Partridge, which occurs from central to northern Portugal and in northwestern Spain; A. r. intercedens (A. E. Brehm, 1857) in southern Portugal and the remaining part of Spain (Figure S1); A. r. rufa (Linnaeus, 1758), in a major part of France and in northwestern Italy.

The Red-legged Partridge was introduced into Corsica (France) in the VI century [11]. First reported as an endemic resource [12], Caccabis rufa corsa was later attributed to the nominate subspecies [9], as confirmed in a recent genomic study [13]. Although A. rufa could have reached the island of Elba 10 km off the Italian northwestern coast during the Pleistocene marine regressions, a late Middle Ages human-mediated translocation from the nearby Corsica was likewise hypothesized [14]. Accordingly, Forcina and co-authors [13] found a strong genetic affinity between the population from the French island and that from Elba. The Red-legged Partridge was also introduced into Mallorca (Balearics) by the end of the XIII century [15]. First reported as A. r. laubmanni [16], this Spanish population was then considered as intermediate between A. r. rufa and A. r. intercedens [17], and ultimately assigned to the latter [18]. Finally, while the most notorious among the introductions carried out into Atlantic islands was that of French A. r. rufa in Great Britain (1673) by the will of King Charles II [19], earlier imports (XVI century) had occurred in both Azores (A. r. hispanica) [20] and Grand Canaria [21]. On the latter, the population was referred to as Caccabis rufa var. australis [22] before being assigned to A. r. intercedens [21].

The Red-legged Partridge was introduced into Madeira and Porto Santo (Figure 1) soon after their discovery by the Portuguese explorer João Gonçalves Zarco in 1419, or in any case before 1450 [23,24,25,26]. Partridges were reported from Madeira by Alvise Ca’ da Mosto, a Venetian nobleman and traveler, in his ‘Account of the voyages to the west coast of Africa by Alvise de Ca’ da Mosto, 1455–1457’, which was first printed in [27]. Later, Fructuoso also [28] quoted [26] reporting the occurrence of these birds on Madeira. The species was confirmed in the early 1700s [29] and subsequently by several ornithologists [21,30]. In the mid-1800s, it was included by Harcourt [31] in the first checklist of the birds of Madeira (cf., [32]), and later referred to as an endemic taxon, Caccabis rufa maderensis [33]. Today, this island population is assigned to A. r. hispanica on a morphological basis only [9] as genetic data on its relatedness to conspecifics from mainland Europe are not available.

We determined the genetic identity of the Red-legged Partridge from Madeira by comparing mitochondrial DNA (mtDNA) sequences obtained from both (Cytochrome-b, Control Region) modern individuals and (Control Region) museum specimens with an already published dataset covering the entire A. rufa Iberian range. The knowledge of the placement of this population within the Red-legged Partridge’s biogeographical frame was deemed important to adequately inform its conservation management, thus avoiding the potential risk of homogenization and reshuffling of island adapted genotypes with non-native ones [34] through supplementation. Particularly, we also investigated the C. r. maderensis syntypes [33] as standardized reference for the Red-legged Partridge of Madeira, ensuring the historical accuracy and reliability of this study.

2. Materials and Methods

2.1. Study Area

The islands of Madeira, Porto Santo, and the Desertas belong to the Archipelago of Madeira, which is included in the Macaronesia Biogeographic Region [35] with the Azores, Selvagens, Canaries, and Cabo Verde (Figure 1). All these volcanic archipelagos harbor around 18,000 native terrestrial species of which 6400 are endemic ([36] and references therein). Madeira is located at 32°74′ N and 16°99′ W and its land surface spans over 758.5 km², with about 2/3 comprised within the limits of a natural park (Figure 1). The island is crossed longitudinally by a sinuous central mountain ridge, with deep valleys departing toward the north and very high cliffs along the coast. In the western part, the ridge flattens out into a large plateau while to the east the central area hosts Pico Ruivo (1862 m), the highest among ten peaks rising over 1400 m. Madeira has a subtropical climate (see also the following paragraph) with Mediterranean summer droughts and winter rain with more humid, colder, and scarcely populated northern slopes, which host the largest primary laurel forest on Earth, and a drier, warmer, and thickly settled southern side with widespread agriculture and plantations [30]. Many endemic taxa occur on the island; among birds, the Trocaz Pigeon (Columba trocaz), Zino’s Petrel (Pterodroma madeira), the Madeira Chaffinch (Fringilla maderensis), and the Madeira Firecrest (Regulus madeirensis) are worthy of mention. Several introduced species can be found as well, including a popular gamebird such as the Red-legged Partridge. As to the latter, the island’s population was estimated to range between 1000 and 5000, breeding pairs in the years 2013–2018, a size larger, for instance, than that of Italy [37].

2.2. Biological Sampling

Partridges were sampled by F.B. in July 2024 in the Parque Ecólogico do Funchal, Pico do Areeiro, and Achada do Teixeira (Figure 1), three areas not only inside the Natural Park of Madeira but also well within the species’ island stronghold according to the data collected by the iNaturalist community. Although partridges can be found almost everywhere on Madeira, they are rarer along the southern coast, west of Paúl da Serra and from Pico do Areeiro to Ponta de São Lourenço, and on the northern slopes (273 A. rufa records gathered between 11 March 2016 and 27 January 2025: exported from https://www.inaturalist.org (accessed on 6 February 2025)). Particularly, the mountainous range of the civil parish of Santo António, São Roque, and Serras do Poiso, along with Paúl da Serra and the entire southern coastal region of Madeira (Figure 1), are the preferred ones by hunters, with Santo António and São Roque also representing the most important areas for A. rufa supplementation. With reference to the three sampling areas, it is worth noting that the Parque Ecólogico do Funchal is mainly characterized by a Mediterranean microclimate, with a long dry season during summer. The vegetation is mostly dry laurel forest, corresponding to the Semele androginae-Apollonietum barbujanae series mixed with introduced species in some parts. The microclimate is temperate in both Pico do Areeiro and Achada do Teixeira, with mist and sometimes strong wind from northeast; the occurrence of hailstone and snow are common during winter. Vegetation is represented by the heather forest, which corresponds to the Polysticho falcinelli-Ericetum arboreae series [38,39].

Fecal pellets were individually housed in empty plastic tubes (no chemicals added); of these, twelve-one per sampling spot within each area, to minimize the risk of investigating partridges from the same covey—were genotyped. A single blood sample collected in Sendim (Bragança, Portugal) in 2003 was also added to the A. rufa sample size available in Pisa (see below; Table S1).

Tissue samples (six slivers of toe pads, two feathers) were also obtained from eight archival Red-legged Partridges originally from Madeira (Figure 1, Table S1): one specimen (ZMB 2000.39417) housed in the Natural History Museum of Berlin (ZMB) and collected on Pico Ruivo by Rudolf von Thanner (1900); both syntypes for Caccabis rufa maderensis (Figure 2) collected 1902–1903 by Ernst Schmitz, one male (NHMW 38.180) and one female (NHMW 38.179) from Ponta do Pargo and Paúl da Serra, respectively, which were provided by Victor Ritter von Tschusi zu Schmidhoffen [33] to the Natural History Museum of Vienna (NHMW); one specimen (NHMW 38.178) from Ribeira Brava (1903, same collector as above); four specimens from the Funchal Natural History Museum (MMF), two of which (MMF 698, MMF 699) were from unknown localities of Madeira and dated 1933, one (MMF 5098) was collected by João Martins in 1955 (Funchal), and another (MMF 41819) by Maria Berta Passos de Gouveia in 1964 (Machico).

2.3. DNA Extraction

We extracted DNA using c.100 mg of material removed from each fecal pellet by means of a sterile disposable razor blade. Then, we used the QIAamp Fast DNA Stool Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions (final elution, 120 μL) and adding two blanks for each session. DNA was extracted from a single blood sample (see above) using the Puregene Tissue Kit, with no modifications with respect to the indications provided by Qiagen and adding one blank in the session in point (final elution, 30 μL).

Extractions from the archival samples were carried out in a physically separated laboratory adhering to ancient DNA protocols throughout all steps and including a strict isolation of both pre-PCR and post-PCR working areas. UV light and 10% bleach were used to sterilize the surfaces of benches and laboratory devices. We used the QIAamp DNA Micro Kit (Qiagen) following the manufacturer’s instructions as in [40] and using up to 5 mg of material obtained from toe pads or the base of the quill (feathers) (all archival samples, final elution: 100 μL). The reliability of each archival DNA extraction was monitored through two blank controls.

2.4. Genetic Analysis and Data Elaboration

We amplified the partial Cytochrome-b gene (Cyt-b, 1092 bp; total length: 1143 bp) and the entire Control Region (CR: 1156 bp) of 12 partridges from Madeira using the semi-nested protocol of [41], which differs from a nested PCR as one of the primers used in the first amplification is also employed in the second one. Hence, in the 2nd PCR, two reactions per gene were prepared to obtain overlapping fragments (Cyt-b: 1st, 575 bp; 2nd, 734 bp; CR: 1st, 621 bp; 2nd, 689 bp). A standard PCR as in [6] was instead carried out for the single individual from Sendim. All PCR products were purified using the Genelute PCR Clean-up Kit (Merck, Darmstadt, Germany) and sequenced in both directions on an ABI 3730 DNA automated sequencer at BioFab (Rome, Italy). We used CLUSTALX v. 2.1 [42]) to align all the 2248 bp-long Cyt-b + CR joint sequences of this study with those available in Pisa (Spain, 43; Portugal, 12: [43]), which were obtained from A. rufa individuals sampled in 22 localities across the Iberian species’ range (12 + 43 + 12 = 67 sequences; Figure S1, Table S1).

As far as the archival samples are concerned, two overlapping (1st: 263 bp; 2nd: 243 bp) fragments of the CR domain I were amplified in two distinct PCR reactions using primers Aru DLCR-F2 and Aru DLCR-R2 of [44] as well as two new internal primers: Aru CRI-FW (5′-CATTAGCCCCATTTCTCCC-3′) and Aru CRI-REV (5′-GTAACCATTCATAGRTTAGGTG-3′). The final 297 bp-long sequence corresponded to the portion comprised between nucleotide positions n. 93 and n. 388 of AJ222740 GenBank sequence [45]. The CR amplicons—obtained using the PCR profile of [44]—were processed in the same manner as the modern samples. We also used additional 64 GenBank records retrieved from a study by Ferrero and colleagues [44], thus including geographic areas that were uncovered by the former set of Iberian A. rufa (33 localities; Figure S1, Table S1). Overall, this alignment comprised 139 CR sequences (12 + 43 + 12 + 64 + 8).

We used DNASP v. 6 [46] to infer haplotypes, which were referred to as ‘H’ and ‘h’ for the Cyt-b + CR and CR dataset, respectively, as well as to compute the haplotype diversity (h ± S.D.) of the Madeiran population (Cyt-b + CR dataset only). For each dataset, haplotypes were used to build a network with the Median-joining method [47] as implemented in NETWORK v. 4.6.1.6 (Fluxus Technology, Colchester, UK).

We ran SMART MODEL SELECTION [48] in PHYML v. 3.0 [49] and we found the TN93 [50] (+G + I for the CR dataset only: α = 0.544; I = 0.725) was the best evolutionary model according to both Akaike (Cyt-b + CR = 7341.8; CR = 1760.1) and Bayesian (Cyt-b + CR = 7862.1; CR = 2159.0) Information Criterion. We used this information to compute the average pairwise genetic distances (d ± S.E., with 10,000 bootstrap replicates for both datasets) among all investigated populations with MEGA X [51]. Then, the distance values were plotted on the first two axes of a Principal Components Analysis (PCA) carried out for each dataset using STATISTICA v. 8.0/W (STATISTICA package, Statsoft Inc., Tulsa, OK, USA).

3. Results

We inferred 41 haplotypes (H) in the Cyt-b + CR dataset (53 variable sites). In the network of Figure 3, haplotypes H1–H28 and H44 defined an A. r. intercedens haplogroup, with the only exceptions of H18 (Zamora) and H20 (León) that were expected to belong to the A. r. hispanica haplogroup (H29–H43) according to their geographic origin (Figure S1). The latter hosted the haplotypes disclosed on Madeira (H35–H37, H40, H41), which were all private to this island. In detail, H35 was the most frequent (41.6%) and widespread along with H41, and it was very close to H38 from Bragança and Asturias; the haplotype diversity was 0.76 ± 0.09.

The groups corresponding to the two Iberian subspecies were also present in the PCA (Figure 4), with the first two components explaining more than 83% of the total genetic variability. The population from Madeira clustered along with those from northern Portugal (n. 32) and northwestern Spain (n. 6, n. 8, and n. 14), with the population from Zamora (n. 7) lying midway between all of the latter and the partridges from the rest of the Iberian Peninsula and the Balearics. In particular, the average genetic distance computed between Madeira and the Iberian populations ranged between 0.0013 and 0.0059, with the closest mainland ones being Bragança (n. 32) and Asturias (n. 8) (Table S2).

We inferred 52 haplotypes (h) in the CR dataset (32 variable sites). All archival specimens dating from 1900 to 1933—included the two C. r. maderensis syntypes (Figure 2)—held the same haplotype (h7), which was shared with modern partridges from all the three areas sampled in the wild on Madeira (Figure 1) and Iberian conspecifics from León (n. 15, n. 16), Valladolid (n. 25), and Álava (n. 22). The two most recent archival specimens—MMF 5098, dated 1955 and MMF 41819, dated 1964—were assigned to haplotype h2 (private to Madeira) and h3, respectively, with the latter being shared with Iberian partridges from Bragança (n. 32), León (n. 15), Zamora (n. 7), Asturias (n. 8), Madrid (n. 12, a single specimen), and the three areas sampled in the wild on Madeira (Figure 5; cf., Table S1). Overall, both modern and archival Madeiran partridges were up to two mutational changes away from each other and/or from the 82.2% (41 out of 47 individuals: Table S1) of the Iberian conspecifics from Bragança (Sendim), Galicia, Lugo, Asturias, Zamora, and León (h1–h9, h13: Figure 5).

In the PCA (Figure S2) the first two components explained more than 73% of the total genetic variability. Modern and archival Madeiran partridges clustered very close to each other and well within the A. r. hispanica populations except for that from Zamora (n. 7), while the A. r. intercedens ones stretched along the first component. The average genetic distance computed between the archival specimens originally from Madeira and the modern Iberian conspecifics ranged between 0.0044 and 0.0254, with the closest mainland populations being those from León (n. 16) and Asturias (n. 8). Finally, the distance (d ± S.E.) was 0.0026 ± 0.0020 between archival and modern partridges of Madeira (Table S3).

All the haplotypes new to this study were deposited in the GenBank (accession codes: PX317556-PX317569, Table S1).

4. Discussion

Small size and relative ease of transport have facilitated human-mediated translocations of the Red-legged Partridge over long distances, this species mainly representing above all—but not exclusively—an important resource as food since the ancient times [14]. Introduced in the Archipelago of Madeira in the XV century, today Alectoris rufa is a regular breeder on the homonymous island [52,53] while it became extinct on Porto Santo in the mid-1800s [25] before being established again in the 1920s [21] using A. r. intercedens founders from Algarve [30].

All investigated partridges from Madeira turned out to hold a maternal lineage (A. rufa) corresponding to their phenotype. Based on a few previous studies [19,54,55], this result could be expected for the archival specimens, whereas it could not be taken for granted when the modern partridges were considered (see below). Despite the small sample size of this study, the lack of A. chukar maternal ancestry represents an important finding. After WWII, A. rufa suffered from an anthropogenic hybridization with the Chukar Partridge aimed at producing more prolific and, as such, more rewarding birds when sold for restocking (for the consequences in the wild, see [56,57]). Although the introgression of allochthonous genes into the genome of A. rufa was recently demonstrated to be much less widespread than previously inferred by many former studies ([13] and references therein), the risk for such an event in Portugal had already been emphasized [58]. These authors reported that about 30,000 captive-bred A. rufa were imported from France and Spain in 1990 (likely, many more in recent years [59]). More importantly, the disclosure of A. chukar mtDNA in phenotypic Red-legged Partridges from northern Portugal was proved by [54]. On the one hand, the lack of A. chukar genes in the partridges from Madeira did not entail that the latter are pure birds, as a nuclear DNA investigation is needed for any reliable conclusion in this respect, especially when the occurrence of past introductions of birds from mainland Europe is considered (see below). On the other hand, this result was a promising one, as the mtDNA (haploid, uniparental) is more prone to introgression than the nuclear DNA (diploid, biparental), also because of selective factors (e.g., thermal adaptation). This was very likely the case for the occurrence of A. chukar maternal genes in A. rufa birds from Elba Island, which, instead, did not turn out to be introgressed using 168,000 genome-wide loci [13] or 11 microsatellites [60].

Captive breeding to supplement the A. rufa populations of Madeira and Porto Santo has been operating at the Forest Station of Casa Velha since 1962, with the local stock from time to time being reinforced with individuals from mainland Europe to prevent its genetic impoverishment; sometimes, local hunting associations were also used to import partridges [61]. Following a modernization of the island farm, the production of partridges increased in the early 2000s, and captive birds were used on Madeira either for restocking in the wild especially across the mountainous range of Santo António and São Roque or sold for hunting competitions. However, between 2007 and 2011, local hunters traded in partridges from mainland Europe; hence, the island’s captive breeding program was devoted to supplementation only in those years (50 A. rufa pairs and 480 birds produced in 2012, [62]). Overall, this management allowed A. rufa to survive on Madeira despite a strong hunting pressure (c. 1000 licenses released in 2017), with the production returning to a high level in 2016 (124 pairs and 2000 birds) [63]. On the contrary, the A. rufa from Porto Santo did not actually require any supplementation over the course of time, and compared to Madeira, both a lower number of hunters (c. 300 licenses in 2019) and a more suitable habitat warranted for a recent flourishing of the species [64].

When the Cyt-b + CR network was considered (Figure 3), we found that all the haplotypes held by partridges originally from the A. r. hispanica and A. r. intercedens range gathered into two distinct groups. This result did not come as a surprise as the genetic divergence of A. rufa from northwestern Iberia was initially reported by [44] and confirmed by [13]. Interestingly, the five haplotypes disclosed on Madeira (H35–H37, H40, H41) were all private to this island and their overall diversity (0.76 ± 0.09, N = 12) was higher than that disclosed by [43] using the same mtDNA markers in A. rufa populations from either larger (Corsica: 0.60 ± 0.08, N = 48) or smaller (Elba: 0.63 ± 0.07, N = 44) islands. The most widespread haplotype (H35) on Madeira was very close to those held by A. r. hispanica representatives from, on the one hand, Bragança (northern Portugal) and, on the other hand, Galicia (type locality for this subspecies, [65]) and Asturias in northwestern Spain. These results agreed either with [25], who suggested northern Portugal as the most credited region as the source for the introduction of the Red-legged Partridge into Madeira, or [66], who literally reported ‘northern Spain’. More broadly, these findings were also in line with a recent study [67] where it was reported that the first human settlers of Madeira (from 1425 on) came mostly from northern Portugal, including historical provinces such as Douro Litoral and Minho, the latter neighboring Galicia. In detail, it is worth noting that haplotypes H18 and H20, which belonged to partridges from Zamora (n. 7) and León (n. 15, Palacios del Sil), were assigned to the A. r. intercedens haplogroup (Figure 3). Although northwestern Spain—compared to rest of the country—is characterized by a lower level of supplementation for hunting purpose [13], these partridges might belong to stocks imported from another region and released into the wild (see also below). The PCA of Figure 4 returned the same picture, with the mainland populations from Bragança (lowest average distance value), Asturias (n. 8), Galicia (n. 6), and León (n. 14, Villafranca del Bierzo) hosting the genetically closest individuals to the partridges from Madeira.

When the CR dataset including both modern and archival partridges was considered, the haplotype network (Figure 5) returned the same result as that conveyed by Figure 3. The four haplotypes (h1–h3, h7) held by both modern and archival Madeiran partridges grouped with those from the very large majority (87%, see Section 3 Results) of A. r. hispanica partridges from northern Portugal and northwestern Spain. This cluster perfectly corresponded to that from northwestern Iberia reported by [44] and included the occurrence of rare individuals from central-eastern Spain, very likely due to restocking practices. The two C. r. maderensis syntypes (haplotype h7) turned out to represent an effective taxonomic reference for the Red-legged Partridge from Madeira, as they reflected the genetic homogeneity existing between the historical and modern island population. Indeed, in the PCA of Figure S2, on the one hand, the populations from northwestern Iberia were the most similar to the archival specimens from Madeira; on the other hand, the latter were closely related to the present-time partridges of the same island. Overall, compared to the Cyt-b + CR dataset (N = 67 sequences, each 2248 bp-long), the larger size (N = 139) of the CR-only dataset did compensate satisfactorily for the shortness of this gene fragment (297 bp-long). A similar choice also worked out very well in a biogeographical investigation [68] of modern and archival Black Francolins (Francolinus francolinus), another medium-sized phasianid, where a shorter portion of the CR (185 bp-long) was used. The sedentary habits of both species as well as the high evolutionary rate of the CR—even when short fragments are considered [69]—represent two possible reasons explaining the effectiveness of the marker in inferring the spatial genetic structure in the two species. On the contrary, in a previous study on modern and archival A. rufa [55], a Cyt-b gene portion (229 bp-long) had failed in assigning a single Madeiran partridge dating from 1906 (FMNH 400948, Field Museum of Natural History, Chicago) to any subspecies, as it shared its haplotype with many individuals from all the existing three.

We are aware that the number of modern and archival A. rufa samples from Madeira was not that high in this study; nonetheless, we did not disclose any significant change in the haplotypic pattern of the local island population over time. The results suggested that the Madeiran population has kept the original genetic distinctiveness typical of the Galician partridge (A. r. hispanica) [13,44], and that the relatively recent importations of partridges from mainland Europe were not significant as hypothesized since [58]. Interestingly, this pattern was also reflected by specimen MMF 41819, which was collected in 1964 in the municipality of Machico, namely a hunting ground not far from the Forest Station where the only A. rufa island farm was operating since 1962 [61].

To conclude, island environments require major attention in the management of the local biota as the small size and naive nature of most populations render them particularly susceptible to extinction [70]. In this respect, we provided an important—albeit preliminary—piece of knowledge for the management of an economically valuable and popular game resource of Madeira. On the one hand, an extended investigation using genome-wide markers and including the only island captive-bred population is needed to receive a more comprehensive genetic picture for the management of the local A. rufa. On the other hand, we recommend not to translocate partridges from both the nearby Porto Santo (and vice versa: [30]) and continental Europe, thus avoiding the risk of introducing genes (at least) from different subspecies, as has, for instance, recently been averted in the Black Francolin hunted on the island of Cyprus [71].