Human Evolution

, Volume 21, Issue 2, pp 167–175

Artificial Occurrence of the Fallow Deer, Dama dama dama (L., 1758), on the Island of Rhodes (Greece): Insight from mtDNA Analysis


    • Dipartimento di Biologia Animale e Genetica “Leo Pardi”Laboratori di Antropologia ed Etnologia, Università di Firenze
  • Alberto Cavallaro
  • Elena Pecchioli
    • CEA, Centro di Ecologia Alpina
  • Cristiano Vernesi
    • CEA, Centro di Ecologia Alpina

DOI: 10.1007/s11598-006-9014-9

Cite this article as:
Masseti, M., Cavallaro, A., Pecchioli, E. et al. Human Evolution (2006) 21: 167. doi:10.1007/s11598-006-9014-9


We investigated the origins of the fallow deer (Dama dama dama) of Rhodes by both morphological and molecular means. Our results show that these deer have homogeneous phenotypic patterns. All specimens fell within the common colour coat variety typical of the wild form. The Rhodian deer appear to be rather small, especially when compared with specimens from central and northern Europe. We then sequenced the HVR-I of 13 deer from Rhodes and compared these sequences with other 31 samples obtained from different European and Anatolian populations of fallow deer. Out of 44 sequences, 23 haploypes were found. When compared to the Turkish and Italian populations, the population of Rhodes revealed lower values of within population genetic diversity. The fallow deer from Rhodes are characterized by an 80-bp mitochondrial DNA (mtDNA) insertion not found elsewhere. As a consequence, all the deer from Rhodes form a tight cluster, distinct from all other fallow deer populations. This uniqueness makes the conservation and management of the Rhodian population particularly urgent.


geneticsD-looppopulation geneticsdiversityconservation


Legend recounts that the fallow deer was introduced to Rhodes from Asia Minor (to wipe out snakes) by the Knights of Saint John of Jerusalem, who conquered the island at the beginning of the fourteenth century [2, 3, 7, 11, 14, 19, 22, 29, 35, 36]. Yet another legend reports that the Turks, for love of a beautiful princess, released these deer on the island only several hundred years ago [35]. Despite these legends, the discovery of osteological remains of the taxon among the Neolithic context (6,000 years b.c.) of Kalythies (cf. [16, 17, 24]) suggests that the introduction of fallow deer occurred much earlier [33]. This Neolithic introduction is not to say that extant deer are simply descendants of prehistoric importations even though an ancient introduction of this breed into the island could explain why the Rhodes population retains wild phenotypical characteristics.

Danford and Alston [10], Festa [14], Wettstein [35] and Masseti [25] published descriptions and dimensions of the fallow deer from Rhodes. According to Festa [14], skulls and mandibles of these deer do not show marked differences from those bred in game parks and reserves throughout Europe. The mayor difference was the narrower palmation of the antlers of the insular cervid. According to Lyddeker [20], narrower palmation of the antlers was a characteristic of the Asia Minor wild fallow deer.

Several authors considered the genetic diversity of the common fallow deer to be very low [12, 18, 27, 28]. The present research on the mitochondrial DNA (mtDNA) of Dama dama dama on samples obtained from Rhodes (Fig. 1) revealed unexpected results. The fallow deer from Rhodes are characterized by a mtDNA insertion not found elsewhere. As a consequence, all the deer from Rhodes form a tight cluster, distinct from all other fallow deer populations. This uniqueness makes the conservation and management of the Rhodian population particularly urgent.
Figure 1

Map showing the provenience of fallow deer specimens from Rhodes and Turkey used in the molecular genetic analysis.

Materials and Methods

The living specimens (N > 100), phenotypically examined in the course of the present study, were mostly bred in the enclosures of the municipality of Rhodes, and a few were occasionally captured wild specimens. To investigate the genetic structure of the fallow deer of Rhodes, we sequenced the HVR-I of 13 individuals and then compared these sequences with other 31 samples obtained from different European and Anatolian populations, including several samples from the Turkish reserves of Düzlerçami (Termessos National Park, Antalya) and Gököva (Mugla), to elucidate their relationships (Fig. 2). We extracted the necessary quantity of DNA from hair tuft, thus avoiding any injuries to the animals. After extraction and purification, HVR-I was amplified by means of polymerase chain reaction (PCR), a technique routinely employed in molecular biology. The DNA sequence was determined through a cycle sequencing reaction, and the vertical electrophoresis was conducted on an automated sequencer. The resulting sequences were then statistically analysed (length = 417 bp).
Figure 2

Network depicting evolutionary relationship between different mtDNA haplotypes from several European populations of fallow deer. Each haplotype is represented by a circle with a diameter directly proportional to the absolute frequency of the haplotype itself. Open circle refers to sequences found in more than one place (but none of these sequences have been found in Rhodes, see Table 1).

Results and Discussion

Phenotypic Pattern

The living specimens, examined in the course of the present, study reveal homogeneous phenotypic patterns. All specimens fell within the common colour coat variety (cf. [10, 14, 22, 24, 35]). None showed the variations towards melanism, partial depigmentation or albinism, which may derive from artificial breeds, such as the semi-domestic populations of Western Europe, selected by man throughout historical times. Thus, the fallow deer of Rhodes are characterized by the colour phenotype regarded as the typical of the wild form (cf. [6, 21]). In summer, the coat color is reddish brown over the head, upper side of the neck, back, flanks and the upper, outer sides of the legs, liberally spattered with white spots on the back and upper flanks. In winter, the coat assumes a darker, duller and greyer brown appearance, and the spots become barely detectable. As already observed by Danford and Alston [10], the Rhodian deer appear to be rather small, especially when compared with specimens from central and northern Europe. In several cases, bucks of over 2 years old do not reach the height of 80 cm at the withers, sometimes not even exceeding 75 cm. They present antlers up to 71–75 cm long.

Population Genetics

Out of 44 sequences, 23 haploypes were found. In Table 1 haplotypes frequencies and sample labels are shown. All but one individual from Rhodes displayed an insertion of 80 bp. This characteristic is absent in the other 31 analysed fallow deer, specimens from Hungary, Italy and Iran (subspecies D. d. mesopotamica). This insertion is undoubtedly a very striking feature, indicative of the distinctiveness of Rhodian population [34], which was already indicated by some analyses carried out on the nuclear DNA, performed through the PCR-modified method called random amplified polymorphic DNA (RAPD) analysis [26].
Table 1

List of mitochondrial DNA control region haplotypes, frequencies and sample labels.



Sample label









Anatolia11, Anatolia12, Anatolia13



Rhodes23, Rhodes10, Rhodes36



Rhodes29 (no insertion)



Rhodes30, Rhodes13, Rhodes22, Rhodes26, Rhodes33, Rhodes35












Italy67, Italy71, Italy75, Italy82



Italy39, Italy96, Italy84, Italy85, Italy89, Italy86






Hungary3, Italy98












Kos10 (Greece)



Kos9 (Greece)



Kos8 (Greece)



DM4, DM5, DM7, DM8










DM: D. d. mesopotamica.

When compared to the Turkish and Italian populations, the Rhodes group revealed lower values of within-population genetic diversity (Table 2), appearing to confirm the hypothesised founder effect. In other words, the population could have been established in Rhodes from a limited number of reproducing individuals, most of them carrying the 80-bp insertion. This kind of event is particularly frequent in the islands and can have important consequences. Effectively, in small isolated populations the differentiation from the other populations can take place even over what is (in evolutionary terms) a relatively short time. That is the reason why endemism is not so infrequent in islands.
Table 2

Indexes of molecular diversity within and between populations.





Rhodes (n = 13)



Anatolia (n = 5)




Italy (n = 19)




On the diagonal: average number of pairwise different within population. Below the diagonal: pairwise Fst values (bold type, p < 0.001). In both cases, the distance method is Tamura and Nei with gamma correction for among-site rate variation (α = 0.32).

To clarify the evolutionary relationships between the Rhodian fallow deer population and the Mediterranean populations described above, we constructed a network of haplotypes (each different mtDNA sequence is a haplotype). The usual methods of inferences such as phylogenetic tree have been originally designed for interspecific comparisons; this means that these methods can not properly deal with the low degree of genetic differentiation and the sharing of ancestral polymorphism between populations that are typical of the analyses at the intraspecific level [9].

The statistical parsimony network [8] we adopted is a method for intraspecific analyses. In this graph, each haplotype is represented by a circle with a diameter directly proportional to the absolute frequency of the haplotype itself, that is, the number of times that haplotype has been found in the sample. The segments linking different haplotypes have length proportional to the number of mutations occurring between different sequences. The longer the segments, the higher the number of mutations separating the haplotypes.

All the individuals from Rhodes (circles with squared), except the one lacking the insertion, form a cluster in the right part of the graph (Fig. 3). This clearly illustrates the distinction of Rhodes from the other European populations. A particularly surprising aspect that emerges is how genetically distant the Rhodian population is from the Anatolian (grey circles – D. dama, Turkey). In spite of the fact that (as previously observed) Rhodes is only a few marine miles from the Turkish mainland, the two groups emerge as extremely divergent at the mtDNA level. Similar results were shown in the maximum likelihood phylogenetic tree (Fig. 3). All the individuals from Rhodes, except the one lacking the insertion (HD5), cluster together as well as the Anatolian samples and the Mesopotamian fallow deer.
Figure 3

Maximum likelihood tree. Numbers above the nodes refer to bootstrap values (1,000 pseudo-replicates). Only bootstrap values over 50% are reported.

The degree of differentiation between Rhodes and the other Mediterranean populations can also be appreciated quantitatively by means of pairwise ϕϕST [13]. This index is derived from the conventional FST and is called the molecular equivalent of FST since it takes into account not only the differences in the frequency of the different alleles among populations (as conventional FST), but also how much the alleles are genetically distant. The ϕϕst values can be comprised between 0, no genetic differentiation and 1, complete genetic diversification. Data in Table 2 shows that the Rhodes population is sharply different from the Italian and Anatolian populations. In particular, the highest value, 0.68 (p < 0.001), concerns Rhodes and Anatolia, thus confirming the divergence highlighted by the network.

Regarding the origin of this diversification, we can only posit hypotheses: the extant fallow deer from Rhodes and from Anatolia diverged a long time ago and have since then remained substantially isolated. As already observed, archaeological evidence indicates the first appearance of fallow deer on Rhodes in Neolithic times, since the 6th millennium b.c.; mtDNA provides us with the confirmation that the deer population of Rhodes was founded by a limited number of ancestors.

Quite unexpectedly, the mtDNA of the descendants of this group does not resemble the mtDNA of the extant specimens from Düzlercami, the Anatolian reserve reputed to be the world’s last stronghold of native populations of this species. Furthermore, on the basis of palaeontological and zooarchaeological knowledge, it could be assumed that Anatolia possibly represented the main source of all European populations of fallow deer after the end of the last glacial episode (cf. [2224]). Consequently, the discovery that the mtDNA of the extant fallow deer of Rhodes is so different from the present Anatolian populations is all the more striking. If we accept the fact that all fallow deer trace back their ancestors to Anatolia, it thus emerges that in the current Rhodian population we retrieve signatures of ancient mtDNA lineages now extinct amongst the sampled Anatolian specimens. This fact makes the Rhodian population particularly important in terms of conservation and management strategies.

Rhodian Deer and Man: Concluding Remarks

Despite intensive hunting of this ungulate throughout history, which shows no signs of abating, even today (cf. [31]), it has been supposed that the taxon survived in Rhodes partly as a result of probable repeated importation from continental Asia Minor since antiquity [1, 4]. It has also been claimed that the last importation of several specimens from Anatolia occurred during the 1970s. This latter supposition is, however, contradicted by the results of the genetic research. Nevertheless, the protection of the fallow deer has recently been guaranteed by government regulation, although the measures are not yet sufficiently effective as regards the scourge of poaching.

In the course of this study, no evidence emerged about the extinction of the fallow deer on Rhodes during the period of Turkish rule (1522–1911), or about its presumed reintroduction by the Italians during the early stages of the Italian occupation (1912–1947), as asserted by Chapman and Chapman [7], following Dicks [11]. In fact, both Wilde [37] and Danford and Alston [10] provided accounts of the deer’s occurrence on the island in the last century of Turkish dominion. In 1880, Charles G. Danford also presented to the Zoological Society of London the complete skeleton of a buck (BMNH: 1880.2.16.2) killed on 22nd December 1878, at Laerma, in the central pine-wooded district of Rhodes where fallow deer were still described as not uncommon [10, 25]. In the early days of the Italian occupation of Rhodes, Festa [14] described fallow deer as still occurring in Laerma, as well as in the areas of Asklipio and Profilia, although their number seemed to have diminished as a result of excessive hunting by the natives. Some years later, Ghigi [15], referring not only to fallow deer but to all wild game, stated categorically that the problem of game reintroduction did not exist on Rhodes, or on any other islands of the Dodecanese. Rather, he suggested that the correct policy lay in protecting the resident wildlife. This statement seems to exclude the desire, at least on the part of contemporary Italian scientists, to reintroduce wild game on the Dodecanese islands.

Tortonese [32] estimated that the number of the wild fallow deer of Rhodes in 1943 was about 2,000, still free-ranging in the interiors of the island. The proof that during the Italian domination, no move was made to import fallow deer originating from Italy can be deduced from the fact that the Rhodian population of the ungulate is made up only of specimens characterised by the wild phenotype, whereas – as we know – about a third of the historic populations of Italian fallow deer is made up of the black variety. Protected during the Italian occupation [4, 15, 30, 35] and by the Greek government since 1969 [1], after the end of the Second World War the Rhodian population constituted an occasional stock for the introduction of fallow deer into various parts of continental Greece. Since the 1960s, fallow deer from Rhodes have also been imported to several Aegean and Eastern Mediterranean islands, such as Lemnos, Lesbos, Crete and Cyprus [5, 24, 25].

The latest techniques for investigating population genetics have shown that the Rhodian fallow deer are very special. They represent an ancient lineage distinct even from the relic populations of the same species sampled in Anatolia, the probable source of the Rhodian stock introduced in Neolithic times. The survival of this population, consequently, becomes charged with a significance that is historic and archaeological, as well as biological, ecological and environmental. As the dynamic testimony of ancient human intervention, which is still available for our evaluation and esteem, its importance has to be considered at the same level as that of a human artefact with all the consequences that such an appreciation brings with it. Unlike several other Mediterranean island populations of remote introduction, however, the Rhodian deer has never been recognised as an endemic taxa, nor has it appeared in the international conservation lists. To conclude, we have to consider that permitting the extinction of this ancient anthropochorous population would, from an ethical and historical point of view, effectively be tantamount to destroying an artistic or archaeological monument.

Copyright information

© Springer Science + Business Media B.V. 2006