Not specified in species description. Robinson and Kloss (1918) substitute the erroneous Bhutan type locality by North Tenasserim, and mention the “type” is at the “British Museum” (NHMUK) but do not specify its catalog number. Moore and Tate (1965) and NHMUK staff suggest the following lectotype: NHMUK 41.1817, field number 213a, an adult male prepared as a mounted skin with a damaged tail & cleaned damaged skull (Fig. 6A).

Figure 6. 

Dorsal and ventral views of the skins of (A) Callosciurus caniceps (NHMUK 41.1817, lectotype); live images of (B) C. caniceps caniceps taken by Natthaphat Chotjuckdikul (Te’) in Bangkok, Thailand; (C) C. caniceps bimaculatus taken by Pattaraporn Vangtal in Railay Bay Beach, Krabi, Thailand; and (D) C. concolor taken by Cheong Weng Chun in Fraser’s Hill, Pahang, Malaysia.

N Tenasserim, Burma [= Tanintharyi province, Myanmar].

A relatively large-sized Callosciurus (average HB = 228.6 mm, PL = 52.6 mm; Table S6) distinguishable from all other relatives by a sharply marked black tail tip along with two alternative pelage combinations corresponding to its subspecies. In the northern part of its distribution, it displays a bright ochraceous/rufus orange on the flanks and dorsum during the dry season and agouti gray in the wet season. The bright orange fades into agouti gray as it approaches the venter, crown, and proximal one-tenth of the tail. The ears are agouti gray, generally with white hairs along the rim, occasionally resembling a tuft (C. c. caniceps; Moore and Tate 1965). In the southern part of its distribution, “the upper parts of the head, neck, all the back and the tail as far as the tip bear a … pelage regularly ringed with ashy and black …” (Temminck 1853), with a rusty reddish suffusion in the side of the neck, and frequently, in the flanks and upper part of the legs; venter coloration is whitish gray, with a red suffusion in the groin area in adult specimens (C. c. bimaculatus and C. c. helvus). Its baculum consists of a shaft and blade; shaft lacks a dorso-lateral expansion at its proximal end; blade base has a reduced lateral expansion at its posterior half, giving it a straight appearance from a lateral view, and a relatively triangular shaped appearance from a dorsal view (Fig. 7); mating call characterized by an average of 8.1 pulses, mean pulse interval of 0.195 seconds, and average frequency modulation of 493 Hz.

Figure 7. 

Baculum comparison between Callosciurus caniceps and C. concolor. A Lateral view of C. caniceps (USNM296998). B Dorsal view of C. caniceps (USNM296998). C Lateral view of C. concolor (USNM283481). D Dorsal view of C. concolor (USNM283481).

A detailed comparison with its closest relative, C. concolor, has been included in the following account. Callosciurus caniceps can be distinguished from all parapatric/sympatric squirrels, including C. concolor, by the combination of gray ventral pelage, lack of a black band in the flank, presence of an abruptly marked black tail tip, and during the dry season, bright ochraceous/rufus orange in flanks and dorsum or rusty reddish suffusion on the side of the neck and groin.

Distributed across mainland Thailand west to the Mekong and up to ca. 19ºN latitude, covering the northern two thirds of peninsular Thailand, and southern Myanmar (Figs 1, S3; south of the Salween River; Moore and Tate 1965). Pelage and craniodental variation suggest its southern distribution limits are in Trang, Surat Thani, and Nakhon Si Thammarat provinces (Fig. S3). However, these provinces were not included in the genetic sampling of Hinckley et al. (2023a), which only included specimens north of 7.99 and south of 7.15 latitude. For a more detailed description of this species’ southern distribution extent, see the account and discussion for C. concolor.

Generalist species found in dry deciduous forest and dry evergreen forest, bamboo forest, secondary forest, parks, gardens, and coconut plantations (Moore and Tate 1965; Kobayashi et al. 2020). In northern Thailand it is only common in bamboo-forests at rather high altitudes (Gyldenstolpe 1914). In Sakaerat Biosphere Reserve in northeastern Thailand, C. caniceps frequently used the ground (34.7%; Kobayashi et al. 2018), which contrasts with little ground use by its sister species’ (ca. 2%; Abdullah et al. 2001; Tamura 1995). In support of this observation, there are 29 records of C. caniceps on the ground from iNaturalist, compared to just two for C. concolor. Further research is required to determine if these behavioral differences are consistent across both species’ ranges. Callosciurus caniceps is frequently found in sympatry with C. finlaysonii, with which it minimizes competition through diet preferences and/or vertical zonation segregation (Kobayashi et al. 2018).

This species has been observed feeding on fruits of Terminalia catappa and Tamarindus indica, and flowers of Triplaris americana in Bangkok parks, and on ten species of plants including Ficus fruits, Sterculia pexa and Acasia seeds, or Mucuna macrocarpa nectar in Sakaerat Biosphere Reserve, and Musa acuminata in Kanchanaburi (Marod et al. 2010; Sommung and Hawkesgood 2016; Kobayashi et al. 2018, 2019, 2020). Fruits and seeds accounted for 71.4% of observed food items in a human-disturbed area of Sakaerat, while bark was the most common food item (ca. 33%) in adjacent undisturbed forests, suggesting a flexible diet (Kobayashi et al. 2018, 2020).

Reproductive information is scarce. A female collected on 11 March 1900 in the Mergui archipelago had two embryos (W.L. Abbott field notes, USNM Mammal Division Archives). No other reproductive information was easily obtained.

Recorded in several protected areas, including Doi Pha Hom Pok, Khao Yai, Kaeng Krachan, Erawan, Nam Tok Tham Sadet, Sirinat, Mu Ko Lanta, Khao Luang, and Hat Chao Mai National Parks, Sakaerat Biosphere Reserve, and Chiang Dao Wildlife Sanctuary (iNaturalist; Table S1). Despite these records, the frequency of detection of C. caniceps was lower in areas with human activity compared to adjacent forests even when human activity and fragmentation impacts were minimal (Kobayashi et al. 2020). This species (sensu lato) is currently considered of least concern. However little is known about its population density, and one study found it to be far less common than a congeneric species, Callosciurus finlaysonii (Kobayashi et al. 2020).

The subspecies C. c. bimaculatus and C. c. caniceps represent two distinct but recently diverged mitochondrial clades, exhibiting limited nuclear differentiation (Hinckley et al. 2023a). Interestingly, both subspecies seem to be present in Koh Tao island. A single Koh Tao “caniceps” could be included in the craniodental PCA and this was clustered with the mainland populations while the two specimens resembling “bimaculatus” (USNM 253434, 253438), actually representing the form “helvus”, showed a distinct morphospace due to their dwarf nature. Unfortunately, the only Koh Tao individuals sequenced in Hinckley et al. (2023a) belonged to the “caniceps” form. Thus, given its distinct morphospace (Fig. 3) and pending molecular evidence, we here consider the pygmy Koh Tao population a valid subspecies: C. caniceps helvus. This subspecies was wrongly synonymized with nominotypical C. caniceps in Moore and Tate (1965) given that the holotype pelage resembles that of C. c. bimaculatus. Previously recognized island subspecies are here synonymized since these generally resemble the closest mainland populations but with slight variations such as a pale tail underside (casensis, and fallax) or a darker (domelicus, fallax) or paler (altinsularis) dorsum.

ZSI9328. Skin and skull, collected or donated by G. Moxon in 1847. This specimen was not cited in the “Annotated Catalogue of the Type Specimens of the Indian Museum” (Khajuria et al. 1977), but the skin is still housed in the National Zoological Collection of the Zoological Survey of India (Uttam Saikia pers. comm.). Other catalog numbers linked to this specimen are: Zoological Survey of India barcode number: ZSI 0000004087; Indian Museum of Kolkata catalog number: “w”.

From the vicinity of Malacca” [Melaka, Malaysia].

A medium-sized Callosciurus (average HB = 218.3 mm, PL = 49.1 mm) that can be distinguished from all other relatives by the combination of the following morphological characters: “reddish suffusion of color occurring on the mid dorsum [that usually extends to the proximal section of the tail] but not on the sides of the neck and body…; absence of a sharply marked off, quite black tip of the tail… [except in northern edge populations; see following sections], and a cool silvery gray color of the venter” (Moore and Tate 1965), that lacks a reddish suffusion in the groin area; cranium with relatively long postorbital processes (LPOP: 3.8–5.8 mm), particularly proportionally to its least interorbital length (LBO: 16.3–18.8 mm); baculum consists of a shaft and blade, shaft has a dorso-lateral (right) expansion at its proximal end, blade base has a large lateral expansion at its posterior half, giving it a sinuous appearance from a lateral view, and a relatively rounded/oval shaped appearance from a dorsal view (Fig. 7); mating call characterized by an average of 17.9 pulses, mean pulse interval of 0.098 seconds, and an average frequency modulation of 1476 Hz.

Callosciurus concolor can be distinguished from its closest relative, C. caniceps, by its smaller size (average HB: 220 vs 235 mm; with the exception of the C. caniceps Koh Tao Island dwarf population, which is even smaller than C. concolor and excluded from this comparison). Additionally, this species exhibits relatively longer postorbital processes and a distinct distribution of ornamented pelage coloration. During the dry season, its dorsum has a reddish suffusion in the midline that generally extends to the most proximal side of the tail, while C. caniceps has a brighter, ochraceous/rufus orange in flanks and dorsum (C. c. caniceps) or a rusty reddish suffusion in the side of the neck and groin, and frequently, in the flanks and upper part of the legs (C. c. bimaculatus; Fig. 6). In the wet season, this reddish suffusion disappears or becomes less conspicuous in both species, but C. caniceps generally still retains a yellow/orange-brown hue. This hue can at least be noted in the side of the neck and groin in C. c. bimaculatus and in a few small patches throughout the dorsum in C. c. caniceps, which greatly contrast with the darker gray ash coloration in the latter. The venter of Callosciurus concolor is silvery gray distinct from the whitish gray (C. c. bimaculatus) or agouti gray (C. c. caniceps). Another diagnostic feature is the absence of a sharply marked off, black tail tip throughout most of C. concolor populations, whereas C. caniceps typically exhibits such feature. Notably, C. concolor populations in Langkawi, Adang, Terutao (⅛ specimens), and Songkhla have a black tail tip that is generally shorter and/or less intensely black than that of C. caniceps, resulting in a less pronounced contrast. However, in Trang, Surat Thani, and Nakhon Si Thammarat, the black tail tip of C. concolor is highly contrasted as in C. caniceps. Examined specimens from the latter two provinces, and all the rain-shade area until Singora have a washed dorsum coloration and possibly lack an ornamented dorsum midline during the breeding season. Thus, the absence of reddish suffusion in the groin area of C. concolor, compared to its presence in C. caniceps bimaculatus, along with the presence or absence of red suffusion on the side of the neck, appear to be the best diagnostic features for distinguishing these species in the contact areas of Surat Thani, Trang, and Nakhon Si Thammarat, and perhaps also in Satun, Phatthalung and Songkhla. However, juvenile and subadult specimens lack this reddish hue, suggesting this feature is diagnostic primarily in adults. The skull of C. concolor is smaller (PL: 46–51.5 vs 49.7–56.2 mm) with a shorter rostrum (SL: 18.0–21.2 vs 20.4–24.4 mm) and interorbital breadth (LBO: 16.3–18.8 vs 18.3–21.7 mm; Fig. S1A), but relatively larger postorbital processes (average LPOP: 4.7 vs 4.8 mm; Fig. S1B) than C. caniceps. As mentioned, the allopatric dwarf population on Koh Tao island has a smaller size than C. concolor and was excluded from these comparisons, but it shows a distinct shape and its morphospace does not overlap with C. concolor, due to the combination of relatively larger postorbital processes and relatively shorter interorbital breadth of C. concolor with regard to the Koh Tao population (Figs 3, S1A, B). Allometric size-controlled statistical analyses only supported significant differences in PL, DL, and MPL between C. concolor and C. caniceps (File S1). Paradoxically, these measurements are the largest, perhaps suggesting an allometric size effect in the data, and that these sister species differ mainly in size, with little shape differentiation.

Baculum morphology provides additional distinguishing characteristics between C. concolor and C. caniceps: Presence vs absence of a dorso-lateral (right) expansion at shaft proximal end in C. caniceps; presence vs absence of a large lateral expansion at blade base posterior half in C. caniceps, giving a sinuous (C. concolor) vs straight (C. caniceps) appearance from a lateral view, and a relatively rounded/oval (C. concolor) vs triangular (C. caniceps) shaped appearance from a dorsal view (Fig. 7).

Finally, C. concolor can be easily distinguished from its closest genetic relative, C. caniceps, by its generally distinct mating call (Fig. 5A; Table S4, correct discrimination rate = 87%). This is characterized by an average of 17.9 pulses, mean pulse interval of 0.098 seconds, and average frequency modulation of 1476 Hz vs an average of 8.1 pulses, mean pulse interval of 0.195 seconds, and average frequency modulation of 493 Hz in C. caniceps (Fig. 5B–D).

It can be easily distinguished from sympatric congeneric C. notatus and C. nigrovittatus by its lack of lateral stripes, and from C. erythraeus by its silvery gray venter, grayish head and limbs which contrast with the remaining dorsum, and darker tail tip vs reddish brown, reddish or agouti venter, homogeneously olive brown colored dorsum, including head, limbs and tail. It can be distinguished from sympatric Sundasciurus robinsoni, S. tahan, and S. tenuis by its larger size, grayish head and limbs, bushier dark-tipped tailed, and thinner and less contrasted pale eye ring, and from S. hippurus by its silvery gray venter coloration with just some black or dark gray/brown in the tail tip vs rusty red venter with entirely black tail in S. hippurus.

Distributed across the Thai-Malay peninsula extending from Melaka to Trang province in the west coast and from central Pahang to Surat Thani in the east coast (Figs 1, S3). Habitat generalist recorded in lowland and hill primary and secondary dipterocarp forests recorded up to ca. 1135 m a.s.l., plantations, and urban/suburban areas (Moore and Tate 1965; Tamura and Yong 1993; Abdullah et al. 2001; Saiful and Nordin 2004). In Ulu Gombak (Malaysia), C. concolor preferred bushy areas with a larger number of small trees, in contrast with C. notatus that showed preference for the opposite (Tamura and Yong 1993). In this reserve, the four species of squirrels use different levels of the habitat: L. insignis uses the ground, C. concolor lower levels, C. notatus middle levels, and C. nigrovittatus the highest levels of trees (Tamura 1995). In the same site, it was shown to feed on 13 species of plants—on fruits (45%), leaves (21%), bark (17%), and flowers (10%)—with “others” comprising the rest of its diet. These consisted mainly of the fruits of Artocarpus elasticus, Piper adancum and Ficus spp., and leaves of Bambusa vulgaris (Abdullah et al. 2001). Other studies in Malaysia based on stomach content examination instead of feeding observations showed that all specimens contained fruit and vegetable matter, and six contained insect remains (n = 9; Harrison 1954, 1961). In Ulu Gombak, the average size of home ranges of males was 2.65 and 3.73 ha, and that of females was 0.79 and 1.54 ha, and females ranges overlap, contrasting with the lack of overlap in other Callosciurus spp. (Tamura 1993; Saiful et al. 2001). Mating bouts were observed from March to July in this site (Tamura 1993). In Selangor, this species has been recorded in sympatry with C. nigrovittatus, C. notatus, C. erythraeus, Sundasciurus tenuis, Sundasciurus hippurus, Lariscus insignis, Rhinosciurus laticaudatus (R. Traub field notes; USNM Mammal Division archives).

Recorded in Labis and Ulu Gombak Forest Reserves, Hala Bala Wildlife Sanctuary, Perlis, Gunong Stong and Selangor State Parks, Tasik Bera Ramsar Site, and Taman Negara National Park (Jayaraj et al. 2013; Ling et al. 2018; William-Dee et al. 2019; Munian et al. 2020; Fauzi et al. 2021; iNaturalist; USNM specimens). It was the most abundant Callosciurus species in a logged hill forest (8.l individuals/ha) but the least abundant one in a primary hill forest, perhaps suggesting a preference for secondary forests (Tamura and Yong 1993; Saiful et al. 2001; Saiful and Nordin 2004). Similarly, it was not recorded in the primary forest of Krau Game Reserve but it was frequently seen in rubber plantations around the reserve (Mackinnon 1978). This species is destructive in plantations of Artocarpus integrifolia and Cocos nucifera in various states of Malaysia (Moore and Tate 1965).

Northwestern edge populations from the islands of Langkawi (named form lancavensis), Telibun [Ko Libong] (named form telibius), Adang (named form adangensis) and Mohea, Nakhon Si Thammarat (named forms mehoeius/mohillius), have a pelage that is somewhat intermediate between C. concolor and C. caniceps bimaculatus, but more similar to the former (This study; Moore and Tate 1965). These have a more sharply marked-off, black tip of the tail than other C. concolor populations (bimaculatus); generally lack the red suffusion in the neck side (concolor), except a single specimen in Lankawi, Ko Terutao and Ko Libong which had some faint yellow C. c. bimaculatus; lack a reddish suffusion in flanks and groin (concolor), and generally have a venter gray coloration that seems intermediate between concolor and bimaculatus (This study; Moore and Tate 1965). Specimens from Langkawi, Adang and Terutao islands were included in the craniodental PCA and are however clearly within the morphospace of C. concolor. Ko Libong museum specimens could not be examined but citizen science pictures suggest these are C. concolor, and/or potential hybrids (Table S3). Remarkably, museum specimens (USNM 83235) and pictures taken 3 km away, in the mainland (Prahmon, Yan Ta Khao District, Hat Chao Mai NP and Mot Tanoi Beach) show a C. caniceps skull and/or pelage. The mouth of the Trang and/or Palian River might represent the northwestern limit for this species at the west coast since a specimen from “Trang” (USNM 86780), which must have been collected in a locality east of this river (according to W.L. Abbott’s field notes all localities sampled during the timeframe in which the specimen was collected were east of this river) and Kao Soi Dao (7.35, 99.86; USNM 258917–8), which is southeast of Trang River, showed C. concolor craniodental morphospace and/or pelage. Potential hybrid specimens are recorded at this river headwaters, in Lay Song Hong (7.83, 99.45; USNM 83495–7), while C. c. bimaculatus with a faintly ornamented pelage side are recorded in Khao Chong (USNM 258916; 7.54, 99.79), at the foothills of the Nakhon Si Thammarat Range (NSTR). This suggests that the Trang and/or Palian Rivers and NSTR might have constituted physical barriers to gene flow to some extent. Similarly, on the east coast, C. caniceps seems to have its southern limit around Surat Thani city and Ban Ta Yai (8.38, 99.87), potentially at the confluence of two major rivers: Ta Pi and Phum Duang. Tha Lo, which is SW of the city, has potential hybrids (Fig. S4). The species C. caniceps seems to be distributed north of these rivers and west of NSTR (e.g., Ban Kok Klap, 8.88, 99.28; Robinson and Kloss 1915), while C. concolor seems to be found to the east of this range (Ban Ta Yai, Nakhon Si Thammarat; Figs 1, S3). Furthermore, specimens (USNM 355078, 355079, 35580, 535161; Fig. S4) collected in this mountain range or at its foothills (Doi Kaeo, Nam Tok Tha Phae, and Ban Tha Phae, at Lan Saka and Chawang districts) have a hybrid pelage (red infusion in mid-line but also in the sides, particularly on the neck, and groin area, or C. concolor venter with C. caniceps dorsum coloration). However, specimens from Ban Nam Tok, Ban Na and Thafa, unknown localities also at Chawang district (USNM 535159, 535163–8), have C. caniceps pelage. Just at 30 km from the potential hybrid population of Nam Tok Tha Phae, and at the east side of NSTR, there is Ban Ta Yai (8.38, 99.87) which has C. concolor pelage. Craniodental data is quite consistent with pelage since localities from Nakhon Si Thammarat and Trang are intermingled with those of N and S groups in the PCA of Hayashida et al. (2007). However, localities from Prachuap Khiri Khan, Chumphon, Krabi and some from Mergui archipelago were also in between the morphospace of C. caniceps and C. concolor, but this study and/or Hinckley et al. (2023a) craniodental, pelage and/or molecular evidence show these populations represent C. caniceps (Hayashida et al. 2006). The addition of the diagnostic postorbital process length variable might have contributed to a better resolution in this study. Similarly, Hayashida et al. (2007) considered specimens with completed molar eruption as adults while this study was more conservative, considering specimens with deciduous or partially erupted premolars as subadults. According to molecular data, the southern limit of C. caniceps is between 6.5–8°N, specimens sequenced further south in Hinckley et al. (2023a) are from Phuket at east coast and Ko Tao at west coast, but sampling at Trang and Nakhon Si Thammarat provinces is lacking.

Holotype not specified in description. Two examined specimens are labeled as cotypes: NHMUK 79.11.21.372, skin and skull, an adult male, also marked as “lectotype”, and NHMUK 79.11.21.373, skin and skull, a young adult marked as “lectoparatype”.

“Assam” [= Assam, India].

A small-sized Tamiops (average HB = 116.5 mm, W = 56.7 g, GLS = 32.7 mm; Table S8; Moore and Tate 1965; Liu et al. 2022) that can be distinguished from all other relatives by the combination of the following characters: relatively weakly striped summer pelage, with outer dark stripes paler than inner dark stripe; narrow (maximum width: 3–5 mm) outer pair of light stripes which are connected to a distinct pale facial stripe; outer light stripes are paler than inner light stripes and have a different width; no traces of rust/orange suffusion in venter (except very rarely in the chest), crown, or nape; absence of a yellow/orange hue in ear tufts; venter hairs have a great extent of gray at their base, which is not entirely covered by the tip lighter coloration, giving the venter an overall dull yellow-buff mixed with gray/brown coloration; golden/yellow tipped tail hairs; relatively large baculum with shaft and blade, shaft lacks a small concavity where the blade is attached, blade lacks ridges, blade attachment orientation is tangent to shaft orientation (Figs 8A, D and 9A, B).

Figure 8. 

Dorsal and ventral views of the skins of (A) Tamiops mcclellandii (NHMUK 79.11.21.372, lectotype), (B) T. barbei barbei (NHMUK 1914.7.8.36, topotype), and (C) T. barbei inconstans (NHMUK 12.7.25.31, holotype); live images of (D) T. mcclellandii taken by Lee Alloway in Kumarkata, Assam, India, and (E) T. barbei taken by Andaman Kaosung in Kaeng Krachan District, Thailand.

Figure 9. 

Baculum comparison between Tamiops barbei and T. mcclellandii. A Lateral (left) view of T. mcclellandii (USNM564453); B ventrolateral view of T. mcclellandii (USNM564453); C lateral (left) view of T. barbei (USNM297043); D ventrolateral view of T. barbei (USNM297043).

It can be externally distinguished from its allopatric relatives Tamiops maritimus and T. mishanica, as well as the allopatric/elevational parapatric T. swinhoei, based on its smaller size and outer pair of light stripes which are connected to a light facial stripe in T. mcclellandii, but just extend to the shoulder in the other species (except in T. swinhoei in which it can be rarely connected with a faint line). Its pelage is also generally harsher, thinner, and shorter vs softer, denser and longer in T. swinhoei (although these features are affected by latitude, elevation and seasonality), and it lacks a yellow/orange hue in its ear tufts and reddish suffusion in crown while these features are frequently present in T. swinhoei. Its facial stripe is distinctive vs diffuse in T. minshanica, and its venter is buff or gray colored, lacking any red hue vs rust colored in T. minshanica (Liu et al. 2022). It can be differentiated from allopatric Tamiops rodolphii based on the following external features: outer light lines are thinner but more distinctive, contrasted and lighter than inner light lines which are less conspicuous vs outer and inner light lines are of equal width and similar brightness, distinctiveness, and contrast with the dorsum in T. rodolphii; outer dark line coloration is homogeneous along the anterior-posterior axis vs outer dark lines grade posterior-anteriorly from black/dark brown to brown/reddish-brown in T. rodolphii; absence of a thin pale brown line across the black mid-dorsal line vs presence of a thin pale brown line across the black mid-dorsal line in T. rodolphii (except in SE Thailand populations); generally duller, venter coloration which goes from buff to yellow (without any trace of red, except in the chest of three Mishmi Hill female specimens) vs generally brighter, orange-yellow to salmon-ochraceous in T. rodolphii. A comparison with Tamiops barbei has been included in the following account, after this species has been formally revalidated.

Distributed across the Eastern Himalayas including Nepal, Sikkim (NE India), Bhutan, Assam and Arunachal Pradesh (NE India), Kachin (Myanmar), and Yingjiang County, north of the Daying River (SW China; Figs 2, S5). It extends southward across the Arakan Range including the Garo, Khasi, Lushai, Mishmi, Naga and Chin Hills (Table S1; iNaturalist; Li et al. 2005; Thapa et al. 2016). Its distribution is possibly limited to the southeast by the Irrawaddy River basin and to the east by the Taping/Daying, Shweli/Longjiang, or Nu Jiang/Salween Rivers since the allopatric T. barbei has been recorded at Maymyo (Mandalay) and Zhenkang (Yunnan), to the west of these rivers (Table S1; iNaturalist; Li et al. 2005). It has been recorded from ca. 170 to 2743 m a.s.l. in tropical and subtropical forest. This species seems to be replaced, perhaps outcompeted, at higher elevations by Tamiops swinhoei in northern Myanmar (Table S1; Ronald Kaulback field notes, NHMUK specimen tags; Moore and Tate 1965). Tamiops mcclellandii rarely comes to the ground, often observed “high up in tall trees, moving in short rushes and then staying motionless, sometimes head downward, often some minutes at a time…never seen one in low bushes” (Lord Cranbrook letter, described in Moore and Tate 1965). Similarly, in Arunachal Pradesh, T. mcclellandii was not recorded on the ground and was shown to mainly forage on bark along tree trunks below 10 m, particularly on bark of the species Kydia calycina, Pterospermum acerifolium, and Amoora wallichii (Datta and Goyal 2008). In northern Myanmar, it was collected at 12 meters up in dead pine tree in dense pine tree and rhododendron forest at 2743 m a.s.l., but also in a yam field surrounded by “light” (possibly secondary) forest at 914 m a.s.l. (Ronald Kaulback field notes, NHMUK specimen tags). Recorded in pairs or small groups often sharing the same tree with Dremomys lokriah (Moore and Tate 1965), and feeding on the bark of Dipterocarpus macrocarpus (iNaturalist). Mating recorded in April in Bhutan (https://www.inaturalist.org/observations/210973847).

Recorded in Neora Valley, Royal Manas, Namdapha, Jigme Singye Wangchuck, Phrumsengla, and Nat Ma Taung National Parks, Buxa and Pakke Tiger Reserves, Eaglenest and Pangolakha Wildlife Sanctuaries, and Cherrapunji-Mawsynram Reserve Forest (Table S1; iNaturalist; Datta and Goyal 2008). Also recorded very close to Khangchendzonga and Phawngpui National Parks (Table S1). This species was recorded in logged forest, but it was significantly less abundant than in nearby unlogged primary forest (Datta and Goyal 2008).

Cyt b pairwise uncorrected genetic distances between populations from Sikkim and the Arakan Range are relatively high (ca. 5–6%) but there is little nuclear differentiation (Hinckley et al. 2023a).

Two syntypes are housed in the National Zoological Collection of the Zoological Survey of India: ZSI9482 and ZSI9483. These represent two unsexed study skins and skulls collected by J. Barbe and have been previously cited by their former Indian Museum catalog numbers: c2 and d2. The third syntype described by Blyth could not be found in this collection (Khajuria et al. 1977). These specimens could not be examined in this study but a topotype was examined (NHMUK 1914.7.8.36).

Zami River, Ye Province, 100 miles south of Moulmein, Tenasserim, Burma [= c. 161 km south of Mawlamyine, Mon, Myanmar; ca. 16.21°N, 97.74°E].

A small-sized Tamiops [average HB = 107–114 mm (T. b. inconstans/T. b. barbei), W = 47 g, GLS = 32 mm; Moore and Tate 1965; Liu et al. 2022] that can be distinguished from all other relatives by the combination of the following features: Outer light stripes are paler than inner light stripes and have a different width; rust/orange suffusion in venter but not in crown or nape; absence of yellow/orange hue in ear tufts; venter hairs have a very small extent of basal gray that is entirely covered by the colored hair tips, giving the venter a homogeneous and bright coloration; white/buff tipped tail hairs (Fig. 8B, C, E). It can also be discriminated based on one of the following combinations of characters, which correspond to its two subspecies. Tamiops barbei barbei: Strongly striped, with outer dark stripes as dark as inner dark stripe; wide outer pair of light stripes (maximum width: 6–9 mm), which are connected to a distinct light facial stripe; relatively small-sized baculum with shaft and blade, shaft has a small concavity where the blade is attached, blade is inserted in this concavity, blade attachment orientation is perpendicular to shaft orientation (Fig. 9C, D). Tamiops barbei inconstans: Faintly striped, with outer dark stripes less dark than inner dark stripe.

In general terms, T. barbei is distinguished from its parapatric relative T. mcclellandii by its more colorful venter and paler tipped tail hairs, and at the turnover area between these species (west and central Indochina), by its strongly striped appearance and longer tail (Figs 8, S2A). Since the dorsal pelage of T. barbei barbei and T. b. inconstans is quite different, and the latter is here considered a candidate species (see comments section), each of these subspecies will be compared separately with T. mcclellandii. Parapatric T. b. barbei and T. mcclellandii can be externally distinguished based on the following pelage features: outer dark stripes as black as inner dark stripe vs paler than inner dark stripe in T. mcclellandii during spring-summer; great vs little contrast between inner light and outer dark stripe pairs in T. mcclellandii; outer pair of light stripes wide (maximum width: 6–9 mm) vs narrow (maximum width: 3–5 mm) in T. mcclellandii; venter gray hair bases are short and entirely covered by the colored hair tips vs venter gray hair bases are long and not entirely covered by colored hair tips in T. mcclellandii, giving T. b. barbei a homogeneous bright yellow/orange ventral coloration vs yellow-buff mixed with gray/brown in T. mcclellandii; white/buff vs golden/yellow tipped tail hairs in T. mcclellandii.

Allopatric T. b. inconstans and T. mcclellandii can be externally distinguished based on less pelage features: venter gray hair bases are short and entirely covered by the colored hair tips vs venter gray hair bases are long and not entirely covered by colored hair tips in T. mcclellandii, giving T. b. inconstans a homogeneous bright orange ventral coloration vs yellow-buff mixed with gray/brown in T. mcclellandii; buff/pale yellow vs golden/yellow tipped tail hairs in T. mcclellandii.

It can also be distinguished from T. mcclellandii by its longer (106–163 mm) vs shorter (75–113 mm) tail in their distribution contact area (west and central Indochina; Fig. S2A). Similarly, cranium morphospace is also more distinct between parapatric T. b. barbei and T. mcclellandii, than between allopatric T. b. inconstans and T. mcclellandii (Figs 4, S2B). More specifically, T. barbei has a deeper rostrum and broader interorbital constriction, but a more flattened braincase than T. mcclellandii (Figs 4, S2B).

Finally, T. b. barbei can also be differentiated from T. mcclellandii through the following baculum characteristics: Presence vs absence of a small concavity where the blade is attached to shaft in T. mcclellandii; blade attachment orientation is perpendicular vs tangent to shaft orientation in T. mcclellandii; shorter (4.70–5.09 mm) vs longer (6.15–6.65 mm) shaft in T. mcclellandii (Fig. 9).

It is currently thought to have an allopatric distribution with its relatives Tamiops maritimus, T. mishanica and T. swinhoei (although T. barbei inconstans might have a sympatric/parapatric elevational distribution with T. swinhoei/maritimus in Yunnan and northern Vietnam) from which it can be externally distinguished based on its outer pair of light stripes, which are connected to the light facial stripe in T. barbei barbei (except in some T. barbei inconstans specimens), but just extend to the shoulder in the other species (except in T. swinhoei in which it is rarely connected with a faint line). Its pelage is also harsher, thinner, and shorter vs softer, denser and longer in T. swinhoei, and it lacks a yellow/orange hue in its ear tufts and reddish suffusion in crown while these features are frequently present in T. swinhoei. Its facial stripe is distinctive vs diffuse in T. minshanica, and its crown is characterized by a dull brown gray vs bright reddish-brown coloration in T. mishanica. It can be differentiated from parapatric T. rodolphii based on the following external features: outer light lines are thinner but more distinctive, contrasted and lighter than inner light lines, which are less conspicuous vs outer and inner light lines are of equal width and similar brightness, distinctiveness, and contrast with the dorsum in T. rodolphii; outer dark lines’ coloration is homogeneous along the anterior-posterior axis vs outer dark lines generally grade posterior-anteriorly from black/dark brown to brown/reddish-brown in T. rodolphii; absence of a thin pale brown line across the black mid-dorsal line vs presence of a thin pale brown line across the black mid-dorsal line in T. rodolphii (except in SE Thailand populations).

Finally, the glandes penis and bacula among T. minshanica, T. swinhoei, T. maritimus, T. barbei barbei, and T. barbei inconstans differ distinctly from each other (Pocock 1923; Liu et al. 2022).

Distributed from the Irrawaddy River to the West to the Northwest provinces of Vietnam to the East (A. E. Balakirev pers. comm.), southern Yunnan (China) ca. 23.5º N to the North, and to Selangor (Malaysia) to the South (Figs 2, S5). In Thailand, the Chao Phraya River Basin might constitute a barrier for this species and its eastern limit, since this species is replaced by T. rodolphii to the east of the Chao Phraya/Pa Sak River (Chon Buri and Chaiyaphum provinces). The only exceptions are several iNaturalist records from Khao Yai National Park, which is just to the East of this river, most of which seem to represent Tamiops barbei. Recorded from sea level to 2130 m a.s.l. (Table S1).

Recorded feeding on cherry blossom flowers/nectar, Mudhuca floribunda, Ficus hirta, Musa acuminata, and Castanopsis acuminatissima (iNaturalist; Rueangket et al. 2019).

Recorded in Khao Sok, Khao Yai, Ko Tarutao, Thai Muang, Kaeng Krachan, Khoa Luang, Phu Suan Sai, Phu Kradueng, Thale Ban, Khao Pu-Khao Ya, Mae Ping, Si Satchanalai and Doi Luang, Doi Inthanon National Parks, Lentang Forest Reserve, Fraser’s Hill and Krau Wildlife Reserves, Chiang Dao, Phu Luang, and Huai Kha Khaeng Wildlife Sanctuaries (iNaturalist; Table S1).

The subspecies Tamiops barbei inconstans might merit species-level recognition. It is well differentiated from T. barbei barbei populations based on mitochondrial DNA, having divergerged ca. 3.54 million years ago with cyt b pairwise uncorrected genetic distances ca. 5% (Hinckley et al. 2023a). Ecologically, it is adapted to subtropical forests instead of monsoon forests/rainforests. Morphologically, it exhibits a highly distinct dorsal pelage, a shorter tail than Indochinese T. b. barbei populations, and a somewhat overlapping but distinct craniodental morphospace. Its baculum also differs from that of nominotypical T. barbei populations (Pocock 1923; Liu et al. 2022). Unfortunately, Hinckley et al. (2023a) were limited by the sampling of only a single T. b. inconstans population and the absence of nuclear data. Future research should expand geographic and genetic sampling to more thoroughly assess whether this subspecies warrants full species recognition.

In T. barbei barbei, morphological sampling in Kra sensu lato complemented and supported the limited molecular evidence in this area of Hinckley et al. (2023a), suggesting a phenotype transition/contact zone between Trang and Hat Sanuk (circa 7.5–12ºN). Molecular evidence showed specimen FMNH82879 from Tenasserim Town (west coast, 12ºN) had south and north haplotypes for different markers. Additionally, FMNH82879 and other specimens from this latitude and all the way south to Trang (west coast, 7.5ºN) also showed an intermediate morphology among north and south groups in the PCA or measurement biplots. However, specimens from Chong Bueng (14.6ºN) show most external measurements in the range of the north group (and T. m. kongensis pelage) while specimens from Nakhon Sri Thammarat and Surat Thani (east coast, circa 8–10ºN) cluster with the south group.