Ungulate taxonomy revisited: the evidence for the splits of G&G

One of the most fiercely debated topics in biology is what constitutes a species. The idea of species is a human-made attempt to capture biological diversity in such a way we can understand it and work with it. Throughout the 20th century the dominant theory was the biological species concept (BSC), which defines species as being reproductively isolated (under natural conditions). The problem with the BSC is how to treat allopatric populations (populations separated without any change of interbreeding) of the same "species". Then it is up to biologists to decide whether the populations are distinct enough to be treated as separate species. As a response to this vagueness, several alternative species concepts arose, with the phylogenetic species concept (PSC) the one that was used most widely. The PSC's key concept is diagnosability, meaning if a population can consistently be told apart by one or multiple characters, it should be considered a separate species. This means that subspecies do not exist under this species concept and any small isolated population can be a separate species if it has at least one trait that is different from all the others. It is this concept that Groves & Grubb followed in their book Ungulate Taxonomy, which caused quite a whirlwind.

Though the PSC sounds appealing, it has the potential of recognizing every small population and even single individuals as separate species, as the diagnosability criterion only needs 1 character that is different, and that could be in the DNA... For more information about the dangers of the PSC with regards to Ungulate Taxonomy read this article:

Are There Really Twice as Many Bovid Species as We Thought? (PDF Download Available)

The fact that Groves and Grubb published a book with their view on how many ungulate species there are is perfectly fine and gives a fruitful discussion. The problem is however that their view was copied without any scrutiny in the Handbook of the Mammals of the World (HMW), but only in 1 chapter. This made it seem that all their splits were valid and as HMW is becoming the standard reference for mammals, it is potentially dangerous when institutions like IUCN would copy their approach based on HMW. After HMW their new taxonomy was further promoted in a Princeton field guide: Bovids of the world by Jose Castello. Again without checking the evidence at hand...

As I am a regular visitor to Africa, the taxonomy of especially antelopes has become particularly confusing and I slowly started to check which antelope species I had seen when using the old taxonomy, which one if G&G are to be believed and additionally what might be the best answer. The ungulate taxonomy used until recently by anyone, was getting outdated, so G&G could have made a major contribution in highlighting the true diversity of ungulates. Unfortunately they were to much in awe of the PSC to make their work helpful in the field... So in this thread I will try to see what the evidence is G&G give for their splits and see whether it holds up, given the data they provide and additional data sources (scientific articles) that are currently available. I am not a taxonomist, but I am a biologist who knows some statistics...

Anyone who has ever read a part of the book Ungulate Taxonomy will have seen it consists of a description of species (groups) with characters named in the text and sometimes aided by tables summarizing measurements on skulls and horns, where they provide sample sizes, means and standard deviation per population/species. There are however quite some things missing from the book:

1) descriptions of the abbreviations they use throughout the book
2) Maps of their proposed new species distributions
3) Any drawings or pictures of characters they deem diagnosable
4) Any results (actual data) of their multivariate statistics, which they use as main tool for analysis
5) They do not perform any tests on their univariate statistics, so they are not checking diagnosability of single characters.
6) They don't give exact locations of where single samples were taken, so one cannot assess whether all samples are along a cline or not, as distributions are often continuous.

Another main problem which has been highlighted a lot already is the often small sample size they use to back-up their claims. There is always natural variation for any single measurement around in biology and one needs a large enough sample size to capture that variation. With a sample size of just a few individuals there is a relatively high chance that you may just have randomly gotten samples that are more extreme than the actual mean of the population you sampled.

In this thread I will go from each species (group) to the next to see how well the evidence holds up and where there might indeed be more diversity than previously assumed. I am not an authority on ungulates, so this is far from a definitive review, more an informed opinion, which is more than many others have managed (e.g. Castello's book).

ARTIODACTYLA


BOVIDAE

Boselaphini

Bovini

Tragelaphini

Neotragini

Procaprini

Raphicerini

Madoquini

Ourebini

Antilopini
Asian Antilopini from here

Oreotragini

Reduncini

Aepycerotini

Cephalophini

Alcelaphini

Hippotragini

Caprini


MOSCHIDAE


CERVIDAE

Muntiacini
Muntiacini continued

Cervini

Capreolini

Alceini

Rangiferini


GIRAFFIDAE


ANTILOCAPRIDAE


TRAGULIDAE


HIPPOPOTAMIDAE


TAYASSUIDAE

CAMELIDAE



PERISSODACTYLA

EQUIDAE


TAPIRIDAE


RHINOCEROTIDAE

 

This is awesome. I'm really looking forward to seeing your results. Are you planning to try and get your findings published anywhere? I imagine this is going to take quite a bit of work and will genuinely be of scientific interest.

 

Great idea for a thread lintworm ! I'm also the opinion that the new "created" species are in many cases overrated and that many of them don't deserve validity !
Only because some skulls are a little larger or the color is a little darker brown isn't IMO a reason to split a species into 2 ( or 3, or 4, or .......) !
Hope you ( and other ZooChatters ) can bring some more light in this complex subject !

 

Thanks Lintworm.

I agree that there seemed to be too many species of bovids in the 'Handbook of Mammals of the World', especially when the same book only listed one species of giraffe. I understand that some species can look very similar externally and can only be distinguished by other criteria, such as calls in galagos. Despite this, I accept that treating individual populations as species seems excessive, especially as the subspecies of desert warthogs were separated by a great distance.

 

Additionally most university libraries (and some larger general ones) will have this book.

That is not the plan currently, for now it is mainly based on personal interest and getting something published is a lot of work and I already have enough papers I need to get published in the coming 1.5 year

So here we go.

Greater kudu

The Greater kudu (Tragelaphus strepsiceros) is a member of the Tragelaphine antelopes, a group that is more closely related to cows than to the other antelopes. On a genus level G&G started their splitting, e.g. putting greater kudu in Strepsiceros, lesser kudu in Ammelaphus, nyala in Nyala and eland in Taurotragus. This splitting is necessary if one wants to keep using the genus Taurotragus for eland, which was originally often done. Genetic studies (Hassanin et al. 2012 Pattern and timing of diversification of Cetartiodactyla (Mammalia, Laurasiatheria), as revealed by a comprehensive analysis of mitochondrial genomes. - PubMed - NCBI & Willows-Munro et al. 2005 Utility of nuclear DNA intron markers at lower taxonomic levels: phylogenetic resolution among nine Tragelaphus spp. - PubMed - NCBI) have however found that Taurotragus belongs inside the Tragelaphus genus and is not a sister to it. Instead of using Tragelaphus for all species in this group (which is followed by Kingdon's Mammals of Africa and Hassanin et al. 2012), G&G split them into several genera. For the sake of simplicity, I will keep using Tragelaphus for all species.

Originally greater kudu has been treated as one species with 3 subspecies:

(source: wikipedia)

T.s. cottoni in Chad/Sudan/ CAR
T.s. chora in the horn of Africa
T.s. strepsiceros in the rest of Eastern and Southern Africa.

G&G have elevated these three subspecies to species status and additionally recognize an isolated in South Africa's Karoo as a separate species ( T. strepsiceros, treating the remainder of the nominate subspecies as T. zambesiensis).

Sample sizes

Skins (Male/Female)

strepsiceros (0/1)
zambesiensis (17/17)
chora (3/3)
cottoni (3/1)

Skulls, horns and teeth (Male/Female)

strepsiceros (2-4* / 1-2)
zambesiensis (12-19 / 9-10)
chora (7-14 / 3)
cottoni (4-7 / 2)

*The sample size can differ within each taxon as not all measurements had the same sample size.

What is apparent from this overview is the very small sample size for all taxa except zambesiensis and chora (males only)

Skins

The main difference in skins between zambesiensis and strepsiceros is supposed to be that strepsiceros are slightly darker.

The main differences differences between the zambesiensis and chora and cottoni are in the number of stripes, the size of the mane

zambesiensis is reported to have 8-11 stripes in males and 5-10 in females. Chora 3-7 in males and 6 in females, where this is 4-8 & 6 respectively in cottoni. So on average zambesiensis has more stripes. The stripes are also more clearly visible in zambesiensis compared to the northern 2 taxa. The mane is absent in female chora and cottoni, but a short mane is present in zambesiensis females. Male chora have only a very small mane, male cottoni and zambesiensis have more of a mane, which is longest in zambesiensis. The mane colour also differs between the 3 taxa.

Horns

There seems to be a difference in the number of turns in the horns:

Number of turns: <2.5. 2.5. >2.5
zambesiensis 5 13 5
chora + cottoni 12 7

(Table 32 G&G)

The northern kudu taxa apparently have on average smaller horns.

Skulls

In their measurements there are no single measures in which strepsiceros is different than zambesiensis. Nor is there any measurement that indicates a difference between cottoni and chora, with one exception that seems to be caused by a huge outlier (cb1 in females).

Skull measurements between zambesiensis and cottoni/chora seem to indicate that several measurements for cottoni/chora are significantly smaller compared to zambesiensis (gt1 & cb1 in males especially and potentially preorbital length as well).

Additional data
There seems to be very little other work done on greater kudu (sub-)species and there is only one genetic study available, which in effect complicates matters further. Nersting & Arctander 2001 (Phylogeography and conservation of impala and greater kudu. - PubMed - NCBI ) looked at the mitochondrial control region of kudu from Nambia to Samburu (Kenya), spanning most of the range of the zambesiensis taxon, but also having one sample from chora. They found a division within zambesiensis between the southwestern and the more northeastern samples, a division that is not reflected in morphological data. Interestingly the 1 chora specimen was quite clearly genetically distinct (though they subsequently argue that it is on subspecies level), but with 1 sample point one cannot draw any conclusions.

Summarizing

G&G recognize four species of Greater kudu, but they do not provide any compelling evidence on which one might split strepsiceros from zambesiensis or chora from cottoni. What their data do however find is a split between the strepsiceros+zambesiensis group and the chora+cottoni group. The greater kudu from the Horn of Africa and Chad have on average smaller horn sizes, less stripes (which are also less clearly visible), slightly smaller skull measurements and less of a mane. Which is backed up by a single genetic sample.

So possibly there are two greater kudu species, one southwestern + central group and one northeastern + northcentral group. Both potential species could in the past have been separated by unfavorable habitat in the past and survived in their own refuge areas. This is not a rare phenomenon in African biogeography, Beisa + Gemsbok and Damara dikdik + Kirk's dikdik are just 2 species that have such a distribution at this moment.

Before anyone can recognize these two species of kudu it would be necessary to have more genetic data supporting this possible split (and also to identify how long both populations were isolated) and it would be good to have more skin samples from the northeastern group. Finally the relict population in the Karoo might warrant further investigation, as they are reproductively isolated from the others.

So there is a chance that in the future we end up with 2 species of Greater kudu or at least two very clearly defined subspecies:
Tragelaphus (s.) strepsiceros (Pallas 1766) and Tragelaphus (s.) chora (Cretzschmar 1826).

Tragelaphus strepsiceros/chora chora

@sebbe67 in Samburu NP, Kenya

Tragelaphus strepsiceros zambesiensis

@lintworm Ruaha NP, Tanzania


@molinea Kruger NP, South Africa

Tragelaphus strepsiceros strepsiceros

@Kudu21 , Addo Elephant NP, South Africa

No photos in the gallery for
Tragelaphus chora cottoni

 

Many thanks for starting this thread; I think that this will develop into one of the most interesting threads on ZooChat.

 

After the Greater kudu it is now time for the Lesser kudu. This is not the sister species of the greater kudu (which are Eland and Mountain nyala). Lesser kudu are however the oldest branch of the Tragelaphine antelopes and are thus sister to all others in the group.

Lesser kudu

Lesser kudu (Tragelaphus imberbis) only occur in Eastern Africa and they have traditionally been placed in two separate subspecies, although it has often been unclear how to identify the subspecies (the description of the southern subspecies was based on 2 female skins) and in the past there has been much confusion over their validity and species distributions. According to Kingdon's Mammals of Africa the following subspecies are recognized:

T. i. imberbis: Ethiopia and Somalia (probably also Sudan and Uganda)
T. i. australis: Kenya & Tanzania


(source:wikipedia)

G&G however recognize two separate species, elevating both subspecies to species status, though changing the distribution, restricting imberbis to East-Central Ethiopia and Northwestern Somalia. Though based on the skull and horn characteristics they imply there is a 3rd taxon in Jubaland (S-Somalia), but they have no skills for that locality.

Sample sizes

Skins (Male/Female)

australis (10/0)
imberbis (12/2)

Skulls/Horns/Teeth

australis (5-8/0*)
imberbis (7-11/0*)
"Jubaland" (5-7/0*)

* Female samples are paired for all regions (n=3)

The sample sizes are relatively ok compared to other species groups treated, but given the infraspecific variation still on the low side.

Samples for imberbis mostly come from the Turkana region in Northwestern Kenya.

Skins
Coats are very variable in terms of color, though overall imberbis males grayer than australis. There is high variation between different australis localities in coat color. Striped in imberbis more spaced than in australis and "rather less marked". australis also has white pasterns, lacking in imberbis.

It must be said that the pictures in Castello's bovids of the world give a very bad impression of the coloration of australis, unfortunately it is far from the only (deliberate?) error made in that book.

Horns

Number of turns 2 2.25 2.5. 3

australis 5 3
imberbis 4 1 4 2
Jubaland 6 1

(Groves & Grubb: page 139-140)

On average australis has slightly less turns in the horn compared to the Jubaland samples, there is no large difference between australis and imberbis.

Skulls

G&G list their three groupings as 100% distinct based on skull and horn characteristics. Based on the data they give in their Table 31 there is absolutely no distinction between australis and imberbis based on univariate analyses and I am very skeptical that it would be the case using a multivariate approach. The Jubaland skulls are on average slightly smaller, but standard deviations overlap quite widely in all characteristics given.

Additional data

There is hardly any additional data available and as far as I have found no work is done on genetics in the wild for Lesser kudu, there are 2 studies on Lesser kudu genetics in captivity, but unfortunately I do not have access to those. Bock et al. 2014 (http://www.readcube.com/articles/10...iley.com&purchase_site_license=LICENSE_DENIED) mention that they found no prominent subspecies structure after sampling captive australis and wild-caught animals from Somalia (taxon unknown to me)

If anyone has access to this article in Zoo Biology, I would be interested to see it.

Summarizing

Based on the data found in G&G there seems to be no reason to treat australis and imberbis as separate species, except if you follow an extremely strict version of the PSC. There is some difference between australis and imberbis in coat pattern, but it would be good to get a more complete sample from all over the range of the Lesser kudu to see whether there might be a cline from grayer males in the north to less gray males in south. The sample from Jubaland is interesting in being slightly smaller, whether this is just a regional variation or hiding the presence of an unknown taxon remains unclear. There is currently no genetic data available from the wild to confirm/contradict recognizing 2 subspecies only, so I do not see any reason to stick to the currently often adopted approach of recognizing both as subspecies, with a remark to stick to the distributions as specified by G&G.

Next stop: Sitatunga

Tragelaphus imberbis imberbis

@Maguari Awash NP, Ethiopia


@Maguari Awash NP, Ethiopia

Tragelaphus imberbis australis

@Tim May Marwell Wildlife, UK


@Tomek Marwell Wildlife, UK

 

Great idea for a thread I agree. Will follow with interest.

 

Sitatunga

Sitatunga (Tragelaphus spekii) is a wetland species, which occurs mainly in Central Africa, south up to the Okavango delta in Botswana and with relict populations in Senegal/Gambia, Lake Chad, Angola and South Sudan. Based mainly on pelage the following subspecies were generally recognized:

T. s. spekii Lake Victoria Basin (Uganda, Rwanda, Burundi, Kenya, Northern Tanzania)
T. s. gratus Congo Basin, West Africa and Sudan
T. s. selousi Bangweulu, Okavango and Zambezi basins (mainly Zambia, Botswana) and a relict population in Angola

Additionally some authors recognize T. s. sylvestris a subspecies only occurring on the Ssese islands in Lake Victoria, this is a sub-species originally described by the fraud Meinhertzhagen and is reportedly a terrestrial subspecies, where sitatunga are normally confined to marshy areas.


(source:wikipedia)

G&G do however split the Sitatunga up in six separate species
T. spekii of the Lake Victoria basin
T. sylvestris of the Ssese islands
T. larkenii of South Sudan
T ugallae of Tanzania (which might be 2 species according to them)
T. gratus of the Congo basin, west Africa and Angola
T. selousi of the Okavango and Bangweulu

So apart from elevating subspecies to species status, they recognize to additional taxa from Sudan and Tanzania.

Sample sizes

Skins (Male/Female)

spekii (13/4)
sylvestris (3/2)
larkenii (4/4)
ugallae (0/2)
gratus (2/2)
selousi (7/6)

Skulls/Horns/Teeth (Males only)

spekii (10-19)
sylvestris (2-3)
larkenii (1-3)
ugallae (1)
gratus:
"albonotatus" (1)
Angola (1)
Cameroon (5-9)
Congo (1-2)
Gabon (1)
Gambia (0-1)

selousi
Bangweulu (3)
Okavango (2-3)



Sample sizes are very small for most taxa, except for spekii and for selousi in the case of skins and gratus when it comes to skulls.


Skins

Selousi is unique in the fact that both sexes have a brown colouration and both sexes have only a hint of stripes, in the other taxa the females have a reddish-brown colouration and the males are darker colored. Sylvestris males have a paler neck and the top of the head is reddish, pre-orbital markings sometimes faint, females without white marks and a dark dorsal stripe. Larkenii males are dark rich brown and both males (4-8) and females (6-7) are striped, though usually faint. Dorsal crest is usually white. For ugallae only 2 female skins were seen, which were quite dissimilar, one was dull yellow-brown with vague white markings, the other reddish with seven clear ill-bordered stripes. Gratus has white lateral stripes and rump spots and is pale-pink brown (the female they mean I guess...). The albonotatus specimen is very similar to gratus but has a larger white frontal chevron, but is otherwise similar and thus retained in gratus. The Angolan gratus specimens are pale to dark brown with hints of 1-2 stripes. Male spekii do sometimes have some white in the crest and females occasionally have rump spots and stripes (as in gratus).


Horns

There is a lot of variation in the horn measurements though very small sample sizes. There is substantial overlap between selousi, ugalleae, spekii larkenii and several of the gratus locations. Sylvestris horns (n=2) are slightly smaller, but easily fall within the variation of spekii, the samples from Angola (n=1) and Gambia (n=1) are larger, but not larger than the largest samples from spekii or selousi.

Skulls

Skull measurements show a lot of overlap and due to the small sample sizes it is difficult to see any real differences, sylvestris skulls are smaller than most other taxa, except that they fall within the range of spekii, the closest taxa in terms of distance. Both the Angola and Gabon specimens look to have slightly larger skulls, but given the small sample sizes that is hard to prove and would only indicate substantial variation within the gratus taxon.


Additional data

There is currently no genetic data available that looks at differences between Sitatunga populations and there are few sources that report variation in pelage patterns of Sitatunga. Many sources however state that single sitatunga populations are extremely variable in pelage patterns as well and in Kingdon's Mammals of Africa it is mentioned based on personal observation of four different scientists that "a comparison of recent field descriptions spanning these basins reveals that these three subspecies (spekii, gratus, selousii) cannot be reliably distinguished on the oft-cited characteristics of pelage color and pattern" (pers. obs. from C. Thouless, J. May, R. Lindholm, pers. comm. E. Stokes).

Summarizing

The section on the coat patterns of the different proposed Sitatunga species may have been confusing, which is true, because they are quite confusing and their does not seem to be a lot of consistence within the proposed species, ugallae has coat descriptions which are just as far apart as spekii and selousii and in gratus some Angolan samples are included which are also more similar to selousii than to the gratus we know (this is the species shown in (most?) zoos). It comes thus of no surprise to me personally that several people have stated that the coat patterns are not a good tool of identification and the sample size used by G&G is much to small to make such wide assumptions. The skull measurements are even less telling as sample sizes are very small for most locations and the one region (Uganda/ spekii) which has a larger number of samples show a wide range in the skull size, so it is very well possible that if one was to have the same sample size for the other taxa, one would find no difference at all between them. As differences that appear to be apparent now (and are cited by G&G and Castello) may be accidental and caused by very small sample sizes.

What is still interesting is the selousii subspecies which in most pictures and also in G&G's data has females the same color as males, but according to the personal observations of other scientists this is apparently not that reliable.

What is supposed to make sylvestris different is that it is a land based taxon, not living in marshy areas, so it's hooves are apparently smaller. I have however seen no data ( the only data on this are the original measurements from Meinhertzhagen himself) on this and pictures seem to be extremely rare of this taxon, so it is currently hard to prove. If it would be the case, one could make a convincing case to treat it as a separate species, because in most other aspects it is extremely similar to spekii, except it's slightly smaller size, which is normal for island taxa...

Overall I see no evidence to split up Sitatunga into 6 species, given that this is a very variable species throughout it's range and the sample sizes and proposed characters by G&G do not warrant such splitting at all. It is even questionable whether subspecies can reliably be recognized, so for now it is maybe best to retain the status quo, with sylvestris as possible 4th valid subspecies. Maybe genetics can give a different picture as the recognized subspecies are occurring in separate drainage basins, so they may have been separated for some time from each other.

Next stop: Bushbuck

Tragelaphus spekii spekii


@jbnbsn99 San Diego Safari Park, USA


@jbnbsn99 San Diego Safari Park, USA

Tragelaphus spekii gratus

@ro6ca66 Whipsnade Zoo, UK


@logroll Axe Valley Bird and Animap Park, UK

no photos in the gallery for
T.s. selousi
T.s. sylvestris

 

This is going to be a long post....

Bushbuck

The Bushbuck (Tragelaphus scriptus) is one of the most widespread antelopes occurring in about 40 African countries. Bushbuck also occur in a wide variety of habitat types, they are only absent from (semi-)arid regions and closed canopy forests, but occur from sea level to 4000 meters altitude... Given their wide distribution and variation in coat patterns a large number of subspecies has been described (up to 46....). There has been some concensus on grouping the west-central subspecies into the "scriptus" group and the southern and east-central in the "sylvaticus" group. This division is based on the fact that the scriptus group are generally brighter colored and are more clearly spotted and striped, whereas the sylvaticus group is more plain in coat patterns.

The following subspecies overview is taken from Kingdon's Mammals of Africa:

"scriptus" group:
T.s. scriptus (including gratus typicus, obscurus) Senegal to Liberia
T.s. phaleratus (including knutsoni, johannae) Cameroon, Gabon, Equa. Guinea, W Congo, W DRC
T.s. bor (including cottoni, dodingae, meridionalis, pictus, punctatus, signatus, uellensis) NE Nigeria, via Lake Chad to NW Uganda & South Sudan
T.s. decula N,S,W Ethiopia and parts of Eritrea

"sylvaticus" group
T.s. sylvaticus southern and eastern coast of South Africa
T.s. roualeynei southeast Africa
T.s. ornatus South-central Africa
T.s. dama (including barkeri, dianae, heterochrous, locorinae, sassae, simplex) Uganda, W-Tanzania, Albertine rift
T.s. delamerei (including eldomae, haywoodi, meruensis, olivaceus, massaicus) from eastern South Africa to central Kenya via Malawi, Tanzania, E Zambia, E Zimbabwe and Mozambique
T.s. fasciatus NE Tanzania, E Kenya, Somalia
T. s. meneliki (including powelli) Ethiopian highlands

The following map from Moodley & Bruford (2007) gives the best indication of which (putative) subspecies occurs where:

We will come back to their research later...

From this subspecies division based on their research G&G recognize the following species:
"scriptus" group
T. scriptus
T. phaleratus
T. bor (including dodingae)

"sylvaticus" group
T. decula
T. sylvaticus (including roualeynei, barkeri, dama, delamerei)
T. meneliki
T. fasciatus
T. ornatus

What is interesting is that contrary to Kingdon, decula is grouped within the "sylvaticus" group.

Sample sizes

Sample sizes for skins are not reported.

Skulls (males only)

scriptus
W Africa (7-9)
Lower Volta (1)
phaleratus (28-39)
bor (22-39)
dodingae (2)
decula (7-8)
powelli (3-4)
meneliki (19-20)
barkeri (8)
dama (40-64)
delamerei (29-46)
fasciatus (9-10)
ornatus (33-45)
sylvaticus (10-25)

Horns (males only)

scriptus
W Africa (2-8)
Lower Volta (2-5)
phaleratus (17-41)
bor (20-42)
dodingae (0-2)
decula (0-7)
powelli (0-4)
meneliki (0-18)
barkeri (3-10)
dama (56-77)
delamerei (27-47)
fasciatus (2-10)
ornatus (32-48)
sylvaticus (8-32)

The reported sample sizes are mostly quite large, the large variation within taxa in horn sample sizes is because horn breadth is hardly measured, whereas horn length and horn span are widely reported. That sample sizes for skins are not reported is somewhat of a problem, as there is a lot of variation in coat patterns. Given the large sample size for skulls and horns, they might have had large sample sizes for skins as well, but of that we cannot be sure.

Skins

scriptus is being described as rick dark rufous with a blackish suffusion, 3-10 distinct stripes and both an upper and lower longitudinal flank band and a circle of white haunch-spots are present. Females are paler than males, but with similar markings. Forelimbs often have a black line right down on the front.

phaleratus is being described as reddish brown without blackish suffusion, except on the withers and especially in females the upper longitudinal band is absent.

bor is being described as more ochery less red in colour, transverse stripes less distinct, no dark suffusion on the withers, upper longitudinal band short or absent, the lower generally broken into spots and streaks. White mark on the throat.

decula is being described as ochery to yellow-brown with ofen a black suffusion on the back, white markings indistinct, 2 longitudinal bands often present in young animals.Front of the forelegs black with white on the front of the knees.

meneliki is being described as dark brown to black in adult males, withonly white on the axillae and occasional white flecks on the haunches and above the hooves. Females are lighter, red with tendency for the white marks to be clearer. Very occasionally traces of white down the forelegs (note: this is a highland taxon)

fasciatus is being described as males dark gray-brown above, gray on the sides, females more yellowish, young animals more rufous. 6 stripes generally distinct, many haunch spots and a broken longitudinal flank band. No black on the crown and the nose, chevron below the eyes in the females.

ornatus is being described as males rich dark rufous, becoming black on the withers, 6-8 white stripes, longitudinal bands reduced to a row of spots, many haunch spots. Females light-brown with fewer stripes. Outer sides of the legs black above the knees, the inner sides white, with a broad black "garter" above the knees and the hock. White stripe from the knees and the hocks to the pasterns.

sylvaticus is being described as older males deep brown to blackish brown with grayish sides, more chestnut above. Younger males more red-brown. White mark on the throat, white spots (usually indistinct) on the face and the haunches, little or no trace of transverse or longitudinal stripes. Females dark yellow-brown to reddish, often with distinct traces of stripes and tend to be lighter on the shoulder and the forelegs.

Skulls

Skulls from phaleratus, scriptus and bor are indistinguishable based on the samples from G&G. Skulls from meneliki and decula (the Ethiopian taxa) tend to be significantly smaller than from the members of the "scriptus" group. Skulls of sylvaticus tend to be larger than of the other taxa, significantly compared to meneliki and decula, but are not really distinct from either fasciatus or ornatus. On average sylvaticus, fasciatus and ornatus skulls are slightly larger than of the "scriptus" group, but it is questionable whether that is significant.

Horns

Horn characteristics are quite similar within the "scriptus" group and not larger then for meneliki and decula as is the case for skull characteristics. Horn characteristics are larger for ornatus, fasciatus and sylvaticus compared to the Ethiopian taxa, decula and meneliki, and the "scriptus" group and those differences are possibly significant.

Additional data

In the past 10 years there have been three studies that looked at mitochondrial DNA of bushbuck across Africa (Moodley & Bruford 2007; Moodley et al. 2009; Hassanin et al. 2012) and all three had the same interesting outcome: bushbuck are not monophyletic. Based on Mtdna it appears that the "scriptus" group and the "sylvaticus" group are separate branches, Moodley et al. (2009) show that the "scriptus" group is the sister to the Nyala, whereas the "sylvaticus" group is sister to Sitatunga and Bongo, both groups are well separated in the cladogram and to call them both bushbuck, it would mean calling all Tragelaphine antelopes bushbuck... The genetic analyses indicate monophyly of all taxa described as species by G&G except for fasciatus which is partly embedded within sylvaticus. The sylvaticus of south Africa are put as sister taxa to the meneliki, rather than with the other sylvaticus samples, this is however not the case with the data used in Moodley et al. (2009).

Whereas other authors and the Mtdna samples retain decula within the "scriptus" group, G&G put them in the "sylvaticus" group based on morphology. G&G propose this is a species of hybrid origin.

Summarizing

Based on morphology and Mtdna there seems to be quite solid evidence for the existence of two separate bushbuck species, which was already recognized in grouping them into an "scriptus" and an "sylvaticus" group. Within both groups G&G recognize several other species, which is mainly due to their strict following of the phylogenetic species concept. Based on morphology one could argue that meneliki is sufficiently different from other bushbuck in the "sylvaticus" group to warrant species status, but this is not reflected in the Mtdna data. Based on skulls and horns it is hard to make any further subdivisions within both groups. Differences in pelage exist but I don't think these differences are large enough for anybody who is not an adept of the PSC to merit those differences species rank... The interesting case is however the decula taxon which might be of hybrid origin and where it should thus be placed. Maybe the use of nuclear dna might shed new light on that.

For now it seems likely that we can recognize two bushbuck species:
Tragelaphus (s.) scriptus (Pallas 1766) and Tragelaphus (s.) sylvaticus (Sparrman 1780)

The molecular studies:
Moodley & Bruford 2007: Molecular Biogeography: Towards an Integrated Framework for Conserving Pan-African Biodiversity

Moodley et al. 2009: Analysis of mitochondrial DNA data reveals non-monophyly in the bushbuck (Tragelaphus scriptus) complex (PDF Download Available)

Hassanin et al. 2012:
http://s3.amazonaws.com/academia.edu.documents/31800584/1-s2.0-S1631069111002800-main.pdf?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=1501416997&Signature=xE7IP6EOkHiB5kA5owvPgcr9mnI=&response-content-disposition=inline; filename=Pattern_and_timing_of_diversification_of.pdf

Next are the remaining Tragelaphine antelopes.


Tragelaphus scriptus scriptus

@ThylacineAlive Gladys Porter Zoo, USA


@ThylacineAlive Gladys Porter Zoo, USA

Tragelaphus scriptus bor

@snowleopard International Wildlife Museum, USA

Tragelaphus sylvaticus sylvaticus


@lintworm , Aberdare Country Club, Kenya



@Hix, Ngorongoro Conservation Area, Tanzania


@Hix, Lake Mburo NP, Uganda


@Newzooboy Kruger NP, South Africa


Tragelaphus sylvaticus meneliki


@lintworm Bale Mountains NP, Ethiopia


@Maguari Bale Mountains NP, Ethiopia

No pictures in the gallery:

Tragelaphus scriptus phaleratus
Tragelaphus sylvaticus decula
Tragelaphus sylvaticus fasciatus
Tragelaphus sylvaticus ornatus

 

Besides the four species (groups) described until now there are 5 other Tragelaphine antelope species, which are are all not split by G&G, in some cases that is a straightforward decision, in others it is slightly less so.

Nyala
The Nyala (Tragelaphus angasii) is a monotypic species without geographical variation.


@Tomek , Zoo Berlin, Germany


@Patrick87 , Zoo Berlin, Germany

Mountain nyala
The Mountain nyala / Gedemsa (Tragelaphus buxtoni) is a monotypic species without geographical variation.


@lintworm , Bale Mountains NP, Ethiopia


@Maguari , Bale Mountains NP, Ethiopia

Bongo
The Bongo (Tragelaphus euryceros) is the sister species of the Sitatunga and hybrids between these species have been recorded in captivity. Traditionally the Bongo has been split in two subspecies, which are geographically isolated from each other:

T.e. euryceros lowland bongo, distributed in the rainforest of West and Central Africa
T.e. isaaci mountain bongo, distributed in montane rainforests in Kenya (and formerly also Uganda)

These two subspecies are often recognized, though Castello's Bovids field guide mentions that there is some debate about there validity. Groves & Grubb do not recognize these two subspecies at all and base this solely on skull and horn characteristics (and small sample sizes), there is some variation here between regions, but nothing alarming, especially considering the small sample sizes. The subspecies are different in color patterns, with the main difference being that male mountain bongo are darker than male lowland bongo. There are no genetic data available that look into the genetic differences between both subspecies. So for now I am happy with retaining both at subspecies level.

Tragelaphus euryceros isaaci


@LaughingDove , Zoo Warsaw, Poland


@Tim May , Whipsnade Zoo, UK

no photos in the gallery of T. euryceros euryceros

Eland

The Eland (Tragelaphus oryx) was traditionally split into three subspecies, allthough there has been some debate about their validity:

T.o. oryx Southern Africa
T.o. livingstonii East-central African woodland areas
T.o. pattersonianus Tanzania northwards

G&G only list oryx and livingstonii as subspecies, based on differences in pelage, though they do not mention whether they actually compared skins themselves. In their analysis of skulls, horns and horns they work with very small sample sizes and there are no apparent differences between regions.

Pelage patterns show a lot of variation, but pattersonianus is darker then the others and with clear stripes, livingstonii is lighter with clear stripes and oryx lacks the white stripes. Mtdna shows two clear lineages, one in Southern and one in Eastern Africa, which have split around 200.000 years ago (Lorenzen et al. 2010 A long-standing Pleistocene refugium in southern Africa and a mosaic of refugia in East Africa: insights from mtDNA and the common eland antelope on JSTOR ). The differences are however as high as might be expected of subspecies in antelopes. The Mtdna shows evidence for only two subspecies, which would support lumping livingstonii and pattersonianus into one subspecies.

So it is likely that Eland should be treated as a single species with two (or maybe three) subspecies.

Tragelaphus oryx oryx

@LaughingDove Entabeni Private Game Reserve, South Africa

Tragelaphus oryx pattersonianus


@Jogy , Maasai Mara, Kenya


@lintworm , Ol Pejeta Conservancy, Kenya


Giant eland
The not-so Giant eland (Tragelaphus derbianus) is the sister species of the Eland, but is not larger, except for the horns. The giant eland has two generally recognized subspecies:

T.d. derbianus Senegal, Mali, Guinea
T.d. gigas Cameroon, CAR, Chad and Sudan

The two subspecies differ in ground color, which is bright rufous in derbianus and sandy in gigas, gigas is also reputed to be slightly larger.

G&G do not recognize any subspecies as in their skull, teeth and horn measurements (very small sample sizes) no differences are apparent. A problem with these measurements is that they appear not to have measurements for derbianus, with only samples from Central Africa and Sudan, except for 2 horns.... Based on comparisons of pictures they do not seem to find differences in coat patterns, but no sample sizes are given...

For now it seems reasonable to keep on recognizing two subspecies of Giant eland, until more data become available.

Tragelaphus derbianus gigas

@ThylacineAlive , San Diego Zoo, USA


@Newzooboy , Johannesburg Zoo, South Africa

No photos in the gallery are available of T.d. derbianus


With this post the first antelope tribe has been covered, the next tribe will be the Reduncini: waterbuck, lechwe, rhebok and reedbuck.

 

Reduncines (Reduncini) are a group consisting of three genera, which originally composed a total of nine species, the majority being closely linked to water. This group is sister to the Oribi, which we will cover later.

Common reedbuck
The Common reedbuck (Redunca arundinum), also known as the Southern reedbuck is the sister species to the Bohor reedbuck (Redunca redunca). G&G note that there is overall less difference between these two species than the traditional arrangement might suggest, they state that R.r. cottoni from Sudan is overall more like a Common reedbuck than a Bohor reedbuck. Annoyingly this is not further explored in their book and based on measurements provided in Kingdon's Mammals of Africa they are easy to separate. Additionally both species overlap in Tanzania and they live next to each other in exactly the same areas (where they are relatively easy to distinguish, even by non-biologists). Note that the Handbook of the Mammals of the World plate is misleading in this case, depicting R. (r.) redunca as the same size as R. (a.) occidentalis, though the latter is the same size as the other "species" of common reedbuck

Back to the Common reedbuck for now. For this species two subspecies are usually recognized:
R.a. arundinum south of the Zambezi river
R.a. occidentalis north of the Zambezi river up to Lake Victoria



As expected G&G elevate both subspecies to species status, though they change the distribution. arundinum is said to occur from South Africa, Zimbabwe, N Namibia, W Zambia (west of the Muchinga escarpment) into SW Tanzania. Occidentalis is said only to occur in Malawi and Zambia, east of the Muchinga escarpment. Strangely both HMW and Castello's bovid field guide recognize both taxa as species, but use the distributions given to them originally as subspecies, so ignoring the distribution as specified in G&G. Annoyingly G&G do not state at any point from where they have seen samples.

Sample sizes

For both skins and skulls no sample sizes are reported for this species (group)

Skins

Arundinum is said to be light grayish fawn, grizzled with brown, with a fulvous "vtinge", especially on the head and neck. Forelegs generally black in front, from the knee to the hoof, hindlimbs frequently with black markings.

Occidentalis is said to be paler an grayer; pale rusty gray on the limbs, tail and body.

Skulls and Horns

Occidentalis has a apparently a longer greatest skull length compared with condylobasal length than arundinum, according to G&G indicating a larger occipital crest; skull relatively narrow, nasals longer and broader (presumably in comparison with arundinum)

No data on horns are given.

Additional data

According to Castello the horns of occidentalis are smaller and less curved compared to arundinum and the former is said to have a dark brown blaze along the bridge of the nose, though this is also present in the provided pictures for arundinum, it is however absent in Bohor reedbuck...

I was not able to find any genetic data

Summarizing

The Common reedbuck case is G&G at it's worst, they do not provide any evidence for what they write, so the details given about skins and skulls can only be treated as anecdotal, which is just not good enough. It also does not help that they give no mention of Southern reedbuck (sub-)specific status in S Tanzania and Mozambique, though that would presumably be occidentalis or the Southern reedbuck from Angola and DRC, which would presumably be arundinum according to them. For now I see absolutely no reason to treat both subspecies as separate species and they also do not give any sound reason why this could potentially be the case. It would be wiser to pay more attention to Mountain reedbuck, but more on that later.

The next antelope will be the Bohor reedbuck.

Redunca arundinum arundinum

@Sarus Crane , Harvard Museum of Natural History


@Nick@Amsterdam Johannesburg Zoo, South Africa

No photos are available of Redunca arundinum occidentalis

 

Bohor reedbuck

As discussed in the previous species, the Bohor reedbuck (Redunca redunca) is the sister species of the Common reedbuck (Redunca arundinum). This species occurs from Senegal to Ethiopia and south into Tanzania, where it overlaps with the Common reedbuck. Based on differences in pelage and horn form up to seven subspecies have been described, though generally only five are recognized:

R.r. redunca Senegal to Togo
R.r. nigeriensis Nigeria, N Cameroon, S Chad, CAR
R.r. cottoni Sudd region in S Sudan
R.r. bohor Ethiopia and the Blue Nile region of Sudan
R.r. wardi E DRC, Uganda, Kenya, Tanzania

Kingdon's Mammals of Africa lists that nigeriensis is sometimes lumped with redunca. In Willem Frost's Antelope of Africa he lists the putative differences in horn form between the subspecies. redunca has rather short horns, that curve strongly forward and incline slightly towards each other. nigeriensis has lower and depressed horns than redunca. bohor has relatively short but larger than in redunca, the horns are less curved and the hooks point forward. cottoni has long and thin horns with a very wide spread, with less hooked tips. wardi has sharply hooked horns.

G&G recognize four species of Bohor reedbuck, recognizing redunca, nigeriensis, cottoni and bohor as separate species, with wardi lumped with bohor. G&G mention an evidently undescribed species from Tanzania which is distinguished by the large size difference between the sexes.

Sample sizes

No sample sizes for skins are reported.

Skulls (males/females)
redunca (5/0)
nigeriensis
Nigeria (7-11/1)
Chad (3-5/1)
cottoni
Sudan W of Nile (13/0)
Sudan E of Nile (10-13/0)
bohor
Ethiopia (11-12/2)
SW Uganda (24-28/3)
E Uganda, W Kenya (10-12/8-9)
E Kenya (2/1)
Tanzania (8-9/2)

Horns (males only)
redunca (5)
nigeriensis
Nigeria (8-10)
Chad (3-5)
cottoni
Sudan W of Nile (12)
Sudan E of Nile (11-12)
bohor
Ethiopia (8-10)
SW Uganda (19-24)
E Uganda, W Kenya (10-11)
E Kenya (1)
Tanzania (7-8)

Sample sizes are good for all taxa, except redunca, when it comes to males, female sample sizes are to small for any real comparison.

Skins

redunca is said to have relatively long dark yellowish fawn hair and no dark markings on the limbs

nigeriensis is said to be light fulvous-fawn, slightly darker along the midline of the back. Paler flanks merge into the white of the underparts. There is a pale dusky stripe down the front of the lower portion of the forelegs. Hairs are short and close.

cottoni has a light color, and has a pale, usually mouse-grey, stripe extending down the front leg.

bohor is grizzled yellowish fawn; dark limb markings sometimes present. Underparts white, sharply defined. G&G note that it is fairly close in general characters to nigeriensis, but separated by cottoni that has different horns.

Skulls

Skull and teeth measurements show that redunca is the smallest taxon, with significantly smaller measurements compared to all other taxa, except the nigeriensis samples from Chad, though these nigeriensis have significantly wider nasals. For the rest there are no apparent differences between the taxa with widely overlapping data between them, differences within taxa are often just as large as differences between taxa.

The notion that Tanzanian Bohor reedbuck have larger size differences is based on two female skulls seems invalid given extremely small sample sizes and large variation, there is indeed a trend that the sex difference is larger than for other localities, but with only 2 females zero conclusions can be drawn and even noting that this might be the case seems to optimistic.

Horns

There is always a lot of variation in horns samples, also within taxa, but there are still some differences apparent. redunca has the smallest horns of all taxa, though ranges vary widely, so I am not sure whether they would be significant, especially as there are only five samples for this taxon. cottoni clearly has longer horns than the other taxa with a larger tip to tip distance as well (on average over twice the distance with redunca, but also larger than other samples on average). Horns from the Ethiopian samples (the bohor subspecies) are larger and wider then horns from other E African localities, except for one sample from E Kenya (wardi). Horns between bohor and nigeriensis are very similar.

Additional data

Unfortunately there has not been any genetic work on Bohor reedbuck taxonomy, as far as I am aware... There has been one study looking at the phylogeny of the Reduncini in general ( Molecular Systematics and Phylogeny of the Reduncini (Artiodactyla: Bovidae) Inferred from the Analysis of Mitochondrial Cytochrome b Gene Sequences), but they do not go to sub-specific level with regards to the reedbuck species.

Summarizing

The Bohor reedbuck is an interesting case as there are some trends visible in the horn shape and size, something that has already been recognized for some time. Differences in pelage seem to be relatively small though... The differences with regards to horns are interesting as they are a classical example of a sexually selected trait. For now I see no compelling reason why the Bohor reedbuck should be split, but given the obvious differences in horns it would be interesting to see genetic data across the range of Bohor reedbuck to see whether there is more hidden variation between regions than morphology alone can show and whether there is a case for elevating some subspecies to species level.

next stop: Mountain reedbuck

Redunca redunca bohor

@lintworm , Bale mountains NP, Ethiopia


@Maguari , Bale Mountains NP, Ethiopia


Redunca redunca wardi

@lintworm , Serengeti NP, Tanzania


@Hix , Serengeti NP, Tanzania

 

Mountain reedbuck

The mountain reedbuck (Redunca fulvorufula) is the 3rd species and probably the most primitive of the reedbucks. Traditionally three subspecies have been recognized, each widely separated from the other two subspecies, each subspecies lives in what is probably a vestige from a glacial period thousands of years ago.

The following subspecies are currently recognized:
R.f. fulvorufula Southern Africa
R.f. chanleri Eastern Africa
R.f. adamauae Nigeria and Cameroon



Not surprisingly each subspecies is treated as a separate species by G&G, additionally they list that it is likely that there are 1 or 2 extra species, one in the Cape region and one in Mongalla (S Sudan).

Sample sizes

No sample sizes for skins are given

Skulls (males/females)
fulvorufula
Cape (1-2/0)
other areas (7-15/2-6)
chanleri (2-10/0-1)
Mongalla (1/0)
adamauae (0-2/0)

Horns
fulvorufula
Cape (2)
other areas (0-14)
chanleri (2-8)
"mongalla" (1)
adamauae (0-2)

It becomes clear that sample sizes are very small for most taxa except fulvorufula. I will ony discuss the male samples as there is only 1 non fulvorufula female sampled.

Skins

fulvorufula is said to be grizzeled grayish fawn, tinged with rufous, especially on the head and neck. Chin, upper throat, underparts and inner limbs are white.

chanleri is said to be ligther and grayer than fulvorufula.

adamauae is described as having a bright reddish colouration, with the red becoming more yellowish on the flanks and white on the belly and internal aspects of upper limb segments. This description seems to be completely taken from Pfeffer (1962).

Skulls

The Cape sample falls completely within the variation of fulvorufula, except that Bas 1 is 3 mm shorter (1.5%) than the smallest fulvorufula measurement (7 mm smaller than the average).

chanleri has significantly smaller measurements compared to fulvorufula for both Gt1 and Bas1 measurements, though the nose is just as broad.

adamauae's two samples are smaller again than chanleri for both biorb and Bas1. The Mongalla sample (1) falls a bit in between chanleri and adamauae, though there is some overlap with the chanleri samples.

Horns

The two Cape samples have smaller horns with a smaller span than the other fulvorufula. chanleri is overall slightly smaller, but still within the variation given for fulvorufula, the same goes for adamauae which 1-2 samples are slightly smaller on average than the chanleri but still fall within the variation.

Additional data

I could not find any data on genetics for Mountain reedbuck.

Summarizing

The fact that G&G propose that there are 2 additional hitherto unrecognized Mountain reedbuck taxa/species, in the Cape and Mongalla, is very hard to defend given the extremely limited sample size and the lack of distinction from the other taxa. The skull and horn data seem to indicate that there are clear(?) size differences between the three recognized taxa, fulvorufula, chanleri, adamauae, but the sample sizes are too small to draw any conclusions from them. They do however show that there might be some significant variation between these three taxa and additional morphological and especially additional genetic work would be important to find out whether the three taxa should be treated as subspecies or species. Also because we do not know their evolutionary history and which factors have exactly shaped the current distribution. Both chanleri and especially adamauae are at risk of extinction, so this research would also be of conservation interest.

Redunca fulvorufula fulvorufula

@Patrick87 , Tierpark Berlin, Germany


@taun , Tierpark Berlin, Germany

Reference:

Pfeffer, Pierre. "Un cobe de montagne propre au Cameroun Redunca fulvorufula adamauae subspecies nova." Mammalia 26.1 (1962): 64-71.

Next stop: Kob

 

Just a small interlude, as I am currently in Ethiopia...

The bovid field guide from Castello uses pictures to illustrate almost every species/subspecies of bovid. Based on these pictures you might be inclined to think that some taxa are very distinctive from their sister taxa based on pelage color. These pictures however need to be treated cautiously as the circumstances during which the pictures were taken do highly influence the color. As an example here 2 pictures from the same day from a Natal red duiker in London Zoo by Tim May (probably even the same individual):





The difference is sunshine and the difference in pelage color is pretty amazing and there are quite some examples (Sitatunga springs to mind) in which Castello's book is not reliable.

And finally some food for thought with an article by Cotterill on Lechwe. Cotterill is from the same PSC school as Colin Groves is, but this article nicely illustrates the methods G&G have used but of which they do not show any results: multivariate analysis. If their book would also contain the results and graphs of those analyses their story could be a lot stronger, if what they claim was distinct using these analyses, was indeed distinct.

The Upemba lechwe, Kobus anselli: An antelope new to science emphasizes the conservation importance of Katanga, Democratic Republic of Congo (PDF Download Available)

 

Thanks for your very intetesting posts in this thread which is one of the most interesting on ZooChat.

I can confirm that the red duiker photos are of the same individual; when I took the second photograph the duiker was standing in very bright sunshine.

 

This is why I like drawn references in ID guides over photographic ones. Because unless the photographer really knows what they are doing in terms of white balance and lighting colour; it can be a real nightmare! For most basic guides photographs work well, and there is a time and place for them as they can often show behaviour or alternate body angles that would be prohibitively expensive to do in drawn art.
But in general for colours, body shape, markings and such its best to have drawn references that show idealistic individuals.

The other problem with photos, esp when one is looking for subtle differences in identification, is individuality. Very few animals are perfect examples of their species; they might have injuries; markings; heck even the body angle/wind/motion can make them appear different to the standard.


An interesting observation of mine though is that many more experienced individuals within bird watching/photography often say that they prefer photographic over drawn guides. I put this down to their greater levels of experience and thus being able to distinguish and interpret good photos better than drawn references. That and some people also find it hard to correspond two different mediums - details that appear clear when drawn but can be muddied or subtle or hard to spot on real examples.

 

Really?!

I agree with all of your points that guides with drawings are superior to photographical ones, but I'm surprised with this last comment.

All serious bird guide books that I have ever had or can think of use drawings and only basic or beginner ones tend to use photos. I also don't know or have ever heard of any serious birders who prefer photographic guides.