Occasional photographic records of butterflies (Lepidoptera, Papilionoidea) in Cambodia: 3, Pursat, Siem Reap, Preah Vihear and Stung Treng Provinces in western, north-western and northern Cambodia

The butterfly (Papilionoidea) fauna of Cambodia is very rich but too insufficiently studied. Results are presented of occasional photographic records of butterflies (Lepidoptera, Papilionoidea) made in 2016–2019 (some also in 2006) along with studies on the Odonata fauna in 67 localities of western and northern provinces of Cambodia: Pursat, Siem Reap, Preah Vihear and Stung Treng. This paper completes the two previous communications of this series devoted to the south-western and eastern provinces of this country. In total, 162 identified and 13 provisionally identified species are listed; 22 species (Miletus ancon, Arhopala agrata, A. alesia, A. allata, A. ammonides, A. atrax, A. aurelia, A. moolaiana, A. silhetensis, Drupadia theda, Anthene licates, Prosotas aluta, Danaus chrysippus, Cirrochroa surya, Doleschallia bisaltide, Athyma ranga, Euthalia recta, Burara oedipodea, Hasora chromus, Pseudocoladenia dan, Koruthalaios sindu, Parnara ganga) are for the first time reported in literature for Cambodia. These and some problematic species are illustrated. Some misidentifications in the two first communications of the series are made; due to reidentification Arhopala camdana and Borbo cinnara are for the first time reported for Cambodia in literature. Acta Biologica Sibirica 6: 293–338 (2020) doi: 10.3897/abs.6.e53770 https://abs.pensoft.net Copyright Oleg E. Kosterin. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. RESEARCH ARTICLE 294    Oleg E. Kosterin / Acta Biologica Sibirica 6: 293–338 (2020)


Introduction
Butterflies are among the most popular insects because of their spectacular wings and general beauty; they also may serve as an indicator of habitat quality. Nevertheless even in spite of this, the butterfly fauna of Cambodia is still far from well studied, mostly because of the sad events of its history in XX century. One-two decades ago this country was still among the most pristine tropical areas but is currently being deforested at a frightened tempo through Chinese and to some extent South Korean investments. Hence the assessment of its biodiversity is especially important now, as most of the natural habitats will soon be lost irreversibly all over the country.
Recent literature on the butterflies of Cambodia is scarce: a popular atlas of the butterflies of SW Cambodia by Woodfield and Murton (2006), reports of two trips by Hiraoki Onodera ( , 2008, results of a stationary study of the Phom Samkos Wildlife Sanctuary by Monastyrskii et al. (2011), the first two communications of this series, based on photographic records (Kosterin 2019a, b), a paper devoted to the genus Hidari Distant, 1886 (Chartier 2019). Yutaka Inayoshi (2019) has summarised the earlier published faunal information on butterflies of Thailand and Indochina the internet site ' A Check List of Butterflies in Indo-China' . It includes all Cambodian reports up to those by Odonera (2007Odonera ( , 2008, the data by Monastyrskii et al. (2011) are only partly included and further data not yet included. A checklist of Cambodian butterflies at present exists in Internet only (Chartier and Kosterin 2020) and is being permanently updated.
In the absence of regular studies of the butterflies of Cambodia, even preliminary and not too precise data appear useful, like photographs taken in nature and identified thanks to the wing pattern, which is quite informative in butterflies and as a rule exhibits diagnostic characters. Of course there are groups where such identification (especially if only one wing side is available) is difficult or impossible, like in the genera Jamides Hübner, [ My trips to Cambodia were focused on dragonflies and damselflies (Odonata) but while studying them I occasionally photographed butterflies and have accumulated quite a number of photographic data. I have already summarised and published those from south-western ) and eastern (Kosterin 2019b) Cambodia. In this third communication of this series I report the photographic data on butterflies from the western (Pursat), north-western (Siem Reap) and northern (Preah Vihear and Stung Treng) provinces of Cambodia, accumulated on six trips made in 2006 (this one only two days long) and 2017-2019.
A considerable amount of data from Siem Reap Province refer to the Phnom Kulen Mts, earlier also studied by Onodera ( , 2008 in 2006 and 2007. Data from Pursat Province mostly concern the Phnom Tumpor Mt. which was thorougly inversigated in 2006 and 2010 by Monastyrskii et al. (2011) as part of the Phnom Samkos Wildlife Sanctuary.
My data from Pursat Province (2019, with a small addition of 2015) are still very scarce. I planned to update them, as well as to complete those on the Phnom Kulen Mts, with one more trip in 2020 but the well known circumstances made this impossible and I find it useful to complete reporting my up to date existing photographic data on butterflies of Cambodia by this communication. Hopefully their accumulation will be continued in future.

Material and methods
The butterflies were photographed in purely natural conditions without any restriction of their freedom and mobility, using Camedia C8080WZ and Canon EOS 350D cameras, the latter with the lens Sigma AF 24-70 mm F2.8 EX DG MACRO. Coordinates were obtained with Garmine eTrex H personal GPS navigator, revised with Google Earth, and provided, in decimal degree format, as intervals (after the decimal point) of actually examined terrain; elevations were retrieved from Google Earth. The dates are given in the dd.mm.year system.
The taxonomic system mostly follows the site ' A Check List of Butterflies in Indo-China' (Inayoshi 2020), but more subfamilies are recognised. Butterflies were identified using the above mentioned site, some taxonomical (Corbet 1941;Evans 1957;Nakamura 2014) or faunal works concerning Thailand (Ek-Amnuay 2006, with corrigenda by Ek-Amnuay et al. 2007), Cambodia (Onodera 2009;2008;2009a), Laos (Onodera 2009(Onodera , 2015 and his unpublished reports kindly provided by the author), Vietnam (Callaghan 2009) and Borneo (Seki et al. 1991), and the internet resource on the butterflies of Tatai Commune of Koh Kong Province of Cambodia by Gerard Chartier (2020), who also offered enormous help in identification of difficult cases.
Most of the photos were identified to species. Provisional identifications are marked with 'cf ' . No subspecies identification was attempted (except for one case), however, most Oriental butterflies have geographically distinct subspecies, so photos were formally attributed to their biogeographically relevant subspecies according to Inayoshi (2020).
Earlier I had uploaded most of photos supporting my two previous communications on Cambodian butterflies (Kosterin 2019a, b) to my internet site (Kosterin 2019c). Now I changed this practice in favour of the Internet portal www. iNaturalist.org, which is easier to handle and includes such an important feature as geolocation. So, all the photographs supporting this communication are available there under the user name @oleg_kosterin, e.g. in the project 'O. 'Kosterin's photos of Cambodian Wildlife'  and can be critically revised. The records presented here have already been taken into account in the dynamic checklist of butterflies of Cambodia (Chartier and Kosterin 2020).
The following widespread morphological abbreviations were used: FW and HW -fore and hind wing, respectively, UPS and UNS -upperside and underside of both wings, respectively, UPF, UPH, UNF and UNH -upperside of fore and hind wing and underside of fore and hind wing, respectively; the numeration of spaces follows the standard British system in which they are numberred according to the dorsal (lower in the common illustration position) bordering vein (see e.g. Ek-Amnuay 2006).

The area O'Som vicinities
O'Som (Veal Veng District, Pursat Province) used to be a desolate village situated on very gentle, moderately elevated (400-600 m a.s.l.) sandstone terrain, actually a low plateau, being part of the Cardamom Mts and covered by vast evergreen rainforest. Until the end of 1990s it was one of the last strongholds of Khmer Rouge, so that for long it was hardly accessible by a muddy road crossing the Cardamoms. Everything has changed recently. A good paved road from Koh Kong City to O'Som was constructed to serve several hydropower stations constructed by Chinese firms, with accompanying small towns, tall dams and vast water reservoirs with partly cleared, partly dead forest at lifeless banks. The largest one, some 6×7 km, is situated north of O'Som. The last patches of rainforest were just cut and were still burning when I visited the village in 2019. The area is now occupied by young plantations among burned stumps and trunks of what recently was a vibrant forest. Yet one of the few remaining population of the Siamese Crocodile (Crocodylus siamensis Schneider, 1801) still exists at a forested river in not far from O'Som, attracting a lot of tourists. Curiously, the nearly pristine rainforest still exists in most of the more southerly Koh Kong Province, and it is the Pursat Province territory which is dramatically devastated. The scarcely disturbed evergreen forest is retained along the Koh Kong Province border, some 6-13 km S of O'Som, which was a target of my day long trip. The natural conditions there were similar to those in Koh Kong Province described in , with gentle hilly terrain clad with continuous forest, a big rapidous river over a sandstone bed, dark slimy forest brooks and banana plantations under tall trees remained from the forest at its border.

Phnom Tumpor Mt. (mostly after Kosterin 2019d)
Most of the Cardamom Mts are formed by sandstone but there is a massif of two twin, partly joined mountains, Phnom Tumpor (or Phnom Kran; 1557 m a.s.l.) and Phnom Dalai in Veal Veang District of Pursat Province of Cambodia, formed by intrusion of the basalt through sandstone. Both mountains are parts of the Phnom Samkos Wildlife Sanctuary. Both are small basalt plateaux with steep slopes, almost table mountains, rising to 1100-1300 m a.s.l. over a gentle terrain of ca 300 m a.s.l. where Pramoui and Tumpor villages are situated. Their upper surfaces are also more or less gentle and covered by a peculiar tall and humid "basalt evergreen forest characterised by very tall trees including giant Ficus, the genera Irvingia, Syzygium and Garcinia, and various representatives of the Lauraceae" (Monastyrsky et al. 2011: 123), which until now is pristine and untouched on Phnom Tumpor, rarely visited by people, and only slightly disturbed on Phnom Dalai.
At the same time, the steep rocky slopes of these mountains, formed mostly by sandstone, are covered with quite different plant associations: dry low forest or open tree stands, scrub, bamboo thickets and, in their middle zone, by the Tennaserim Pine (Pinus latteri Mason) zone on sandstone rocks. In the pine zone on the slopes of both mountains an endemic pitcher plant species occurs, Nepenthes holdeni Mey (Mey 2010). All types of the slope vegetation are regularly subject to low-level fires. At the time of my visit the slopes Tumpor were completely dry and largely burnt out without fresh leaves -almost totally at lower elevations, moderately at medium elevations and scarcely at the top. The upper parts of eastern spur crests bore fragments of taller forest, some scrub and tall Poaceae; burnt areas were infrequent. The mountain top surface is a different world of humid, tall and lush forest on a thick soil and slightly undulating terrain. At the time of our visit, the O'Gran Rivulet had a very weak current -in shallow places like a small brook, but actually comprised a chain of the so-called 'anlung' -deep and broad reaches, connected by short stretches of brook seeping through small stones. Where the O'Gran approaches the mountain slope it flows over a 'staircase' of large rocks and forms two small waterfalls with deep pools beneath and finally fell to the abyss; this rocky part was the only small open area examined.
Thus the upper forested surfaces of both mountains offer a kind of isolated, 'island' habitat, a 'lost world' . In 2000, there still were elephants and tigers on both mountains but both had disappeared from Tumpor by 2006; now the tigers have disappeared from the whole of Cambodia while elephants might still be present on Phnom Dalai (J. Holden pers. comm.).

Phnom Kulen Mts
Phnom Kulen Mts are situated in the northern part of Siem Reap Province and are two twin pear-shaped sandstone plateaux extended from NW to SE and directed by their narrow ends towards each other, with a 1.5 km wide gap between them, through which National Road 67 goes from Siem Reap to Anlong Veng. The NW plateau has the dimensions of 20×11 km, the SE plateau 28×15 km. They have nearly flat, moderately elevated (from ca 200 to 487 m a.s.l.) surfaces but remarkably steep slopes, however the NW plateau has a branchy, lowered but yet rocky western spur. Since the plateaux get at least rare small rains even in the dry season, the natural vegetation of the top surface is evergreen rainforest and that of the slopes is sparse deciduous forest. Both in general appearance and in biogeographical sense Phnom Kulen is a forest island among vast open landscapes -deforested farmland to the south and open low deciduous forests of Dipterocarpus intricatus Dyer to the north. (There is no surprise that Phnom Kulen also served among the last Khmer Rouge strongholds.) The SE Plateau is drained by a considerable river known (more or less changing each other from upstream to downstream) under the names O'Dar, Prey Thom, Chup Preah and Siem Reap. It has a famous tall waterfall being a good butterfly place. There are many smaller rivers and brooks at both plateaux. In the SE part of the SE plateau there are some areas of flat sandstone outcrops with sparse savannah-like vegetation (Khmer 'veal'), the largest situated to the east of Thmei village. There is a remarkable 'floating forest' at Tnal Mareh terrain, where a branchy stream flows through a swamped forest with trees producing numerous pointed pneumatophores, strongly resembling mangroves. A similar habitat was said to exist also at Tmar Truonh village. However, it is unclear if these habitats are fully natural or formed in the bottoms of the former big water reservoirs after them being destroyed (at least a thousand years ago Tnal Mareh served as a clay quarry for a ceramic factory).
The SE plateau has been being densely inhabited for not less than a thousand years and always was a sacred place. There in 802 Jayavarman II proclaimed the independent Khmer Empire (Kambujadesa) with the capital of Mahendraparvata (the same name being applied to the mountains). In later decades he left it in favour of Hariharalaya at Lake Tonle Sap. Later Phnom Kulen served a source of stone blocks for construction of the famous temples of the Angkorian complex, which were transported from there to the Lake Tonle Sap vicinities via impressive artificial channels. Phnom Kulen harbours remarkable monuments of the pre-Angkorian period: thousands of lingas carved (to provide fertility downstream) in the bottom of the Siem Reap River, the biggest in Cambodia Buddah statue called Preah Ang Thom, two ancient elephants carved from rocks, Damrei Krab and Damrei Sras, and quite a number of pre-Angkorian temple ruins of different state of preservation. At the Rong Chen Archaeological Protected Area a ceramic factory had been launched along with the capital and existed unti it moved to the lowland, so that the ground under the forest is presently full of broken ornamental ceramics. There were not less than four big artificial water reservoirs (Khmer 'baray'): Ta Penh, Tnal Dach (the present name; 'broken dam' in Khmer), Srae Tbong and Tnal Mareh. However, their dams were destroyed by American bombing in the course of the so-called Operation Menu in 1969-1970, so that three of them have no more water and only the dam and reservoir of Tnal Dach was restored in 2012. The traditional land use on the SE Phnom Kulen plateau is 'circle agriculture' (a kind of slash-and-burn agriculture): the forest is cleared and burnt in a rounded territory some 150-200 m in diameter, crops are grown, then the circle is abandoned and re-used each 6-7 years. At present most of the plateau surface, as viewed from above, is a mosaic of adjacent cleared circles separated with narrow borders of remaining forest trees. In 2003 cashew was introduced and since 2008-2009 cashew growing appeared the predominant land use. Only some areas of pristine forest are still protected, like Preah Thom (Khmer 'big forest') Community Protected Area (ca 5×3 km) at the Preah Ang Thom Buddha and the waterfall. Presently a good paved road is being constructed through the SE plateau, said to serve for construction of a town of luxuriant villas, naturally for Chinese money.
For some reason the NW plateau remains unexplored and undisturbed still being covered with vibrant pristine rainforest. However, the Kbal Spean (Khmer 'head bridge') River in its SE part is famous for the linga carvings in its bottom, even more impressive than those on the SE Plateau. There is a convenient, broad and clean path going uphill through the evergreen forest from the Apsara Foundation checkpoint to Kbal Spean, which offers a great variety of butterflies to observe. The lowering wester spur of this plateau has high impressive rocks but is scarce in water.

Lake Tonle Sap
In its headwaters, the great Mekong River collects a big share of moisture brought by the summer monsoon to the Sino-Tibetan Mountains being the eastern margin of the huge Tibet Plateau. Having left these mountains, the Mekong River flows to the south through southern China, provides the border between Myanmar and Laos, enters Laos, provides the border between Laos and Thailand, re-enters Laos, enters Cambodia and then enters Vietnam where it at last joins the Gulf of Siam of the South Chinese Sea. Being the seventh longest (4,350 km) river in Asia, with the catchment area of 795,000 km3, it has been carrying a huge amount of sediments which formed the alluvial Cambodian Lowland which once was a bay of the Gulf, so that in its lower reaches Mekong has to pass through its own flat alluvium. This lowland is situated almost at sea level and is so flat that Mekong at its higher levels fails to discharge all its water to the sea. The excess water is therefore redirected and stored in the natural reservoir of Lake Tonle Sap (Khmer 'big lake of fresh water') situated in the centre of the Cambodian Lowland. It is connected with Mekong by the Tonle Sap River which joins it at the city of Phnom Penh, the capital of Cambodia. During the summer monsoon, from May to September, this river flows from Mekong to fill Lake Tonle Sap. When the Mekong level decreases, the current reverses and the water stored in Lake Tonle Sap starts to flow back to Mekong, to be at last discharged to the sea. Hence the Tonle Sap River changes its direction twice a year, approximately in late May-early June and late October-early November, and the Cambodian New Year celebrated on 14-16 th of April is associated with its spring turn (although currently takes place somewhat ahead of this). As a result, Lake Tonle Sap cyclically changes its length of 160 km, the area of 2,500 km2, the volume of 1 km3, and the average depth of 1 m in May to a length of 250 km, the area of 16,000 km2, the volume of 80 km3, and an average depth of 6-9 m in October (Mekong River Commission 2005). The lake is also fed by its NW tributary, the Stung Sangkae (or Sankar) River. Soundings have shown that the lake bottom is perfectly flat without any relief, with the deepest point situating just 2.4 m above sea level (E. Smith, pers. comm.). In these respects, Tonle Sap is somewhat analogous to the Amazon River or even the epicontinental seas of the Mesosoic. Nevertheless it is fairly young, as formed just 6-8 thousand years ago, that is rather shortly after the Holocene onset, obviously following the increase of humidity at higher latitudes and the volume of the Mekong flow.
Tonle Sap is the largest freshwater lake in Southeast Asia and one of the most productive freshwater ecosystems in the world. Like the Nile in Egypt, this huge fluctuating water body full of fertile sediments has always been the source of living resources, through rice farming and fishing (the latter practiced by Cham rather than Khmer people), for 1.2-1.3 million people populating its banks (but in the fishing season lasting from October to June, their number increases to 1.4-1.6 million). This was a prerequisite of the rise of the famous Angkorian Civilisation in IX century, whose capital was the largest urban centre in the world in XI-XIII centuries. On the flatland surrounding the lake, several embankments go parallel to the water front, so that when the lake shrinks, shallow water (partly supplied also with channels from outside the lake) is retained at their inland side and is subsequently used for watering rice fields at their lake-faced side. As a rule these shallow 'reserves' are covered with thick but low temporarily inundated scrub, but at least a huge area (some 3×2 km) SW of Siem Reap City, bordered by a road embankment going from Phnom Krom village to Sambuor village, is a deeper and permanent lotus/water hyacinth swamp.
The natural vegetation of Lake Tonle Sap banks was the peculiar, temporarily flooded forest composed mostly of Barringtonia acutangula (L.) Gaertn., Dyospiros sp. and Terminalia cambodiana Gagnep. (L. Everaere pers. comm.). This forest has gone from the banks because of human overexploitation lasting for at least one thousand year but remains at the lowermost reaches and the mouth of the Sangkae River, along the southern bank and in certain refugia on the northern bank, e.g. in a popular touristic site of the stilted villages of Kampong Pluk.
However peculiar the Tonle Sap ecosystem is, it is currently far from its pristine state not only because of deforestation, extermination of large animals and overexploitation of natural resources but also, if not mainly, due to the prolonged history of alien species invasion. Thus, large areas of the shallow water surface in inundated forest and scrub is covered by carpets of Water Hyacinth (Eichhornia crassipes (Mart.) Solms), the invasive species of the Amazonian origin, which produces enormous biomass and comprises an overwhelming majority of floating vegetation. Among the water hyacinth, rosettes are frequent of the Water Cabbage (Pistia stratiotes L.), another species of probably an American origin. The inner zone of seasonably flooded scrub facing the lake bank is largely formed by the 'Giant Sensitive Tree' (Mimosa pigra L.), a tall semiaquatic spiny bush of a Mexican origin, and side-by-side with it Giant Cane (Arundo donax L.) emerges from water, whose indigenous range probably was in the Mediterranean and Near East, while both are interspersed with floating water hyacinth. Hence most of the bank vegetation below tree level is currently formed by alien species. The actual alteration of the Tonle Sap ecosystem by invasive species may be much more profound than is noticeable by superficial observations.
There is a furiously debated plan to construct a large hydropower dam across Mekong in Kratie Province. It will regulate Mekong and is expected to stop the Tonle Sap level cycle, thus destroying its unique ecosystem and depriving about million people from living resources.
The vast area between Lake Tonle Sap and Phnom Kulen Mts (37-50 km apart) is a farmland: mostly rice fields, some pastures, secondary growth, a lot of picturesque Palmyra Sugar Palm (Borassus flabellifer L.) growing semi-naturally (but having appeared in Cambgodia only ca a thousand years ago). At the same time tall semi-evergreen forest is retained in the territories of the famous temples of the Angkorian complex.

Prey Long Forest
The Cambodian Lowland is a great extension of the lowermost valley of the Mekong River and is filled with its sediments. Once covered by lowland tropical forest, long ago it was mostly cleared and converted into arable land. Presently most of the lowland is almost infinite rice field with scarce trees grown for purpose. Nevertheless, large areas of pristine rainforest have been persisting there until present. The largest of them, among the largest lowland tropical forest in the Old World, Prey Long Forest (or Prey Lang; 'our forest' in the language of local Kuy minority (Turreira-Garcia et al. 2017)) is situated between Lake Tonle Sap and the Mekong River and shared by Preah Vihear (west), Stung Treng (east), Kratie (south-east) and Kampong Thom (south) Provinces. Prey Long Forest extends for ca 73 km both from north to south (ca 12.9-13.6 N) and from west to east (ca 105.3-106.0 E) at elevations 70-170 m a.s.l., occupying 530,000 ha (Turreira-Garcia et al. 2017). It is being furiously logged in Kampong Thom Province, in spite of its status of a wildlife sanctuary. Since lowland wood is cheap to log and transport, persistence of Prey Long Forest was a matter of absence of roads, so that as soon as a road appears the forest is gone. Because of the same absence of roads, it remained totally unexplored scientifically until very recently. Actually it was discovered for science by a 2004 expedition by J. Andrew McDonald (2004).
According to McDonald (2004), the most widespread plant community of Prey Long is (i) primary tall semi-evergreen dipterocarp forest (dominated by Dipterocarpus, Irvingia and Syzigium spp.); other vegetation types occupying less areas are (ii) short semi-evergreen dipterocarp forest, (iii) short deciduous forest (dominated mostly by Dipterocarpus intricatus), (iv) tall sralao (Lagerstroemia spp.) forests, (v) short riparian and Melaleuca forest, (vi) deciduous swamp forest (at Lake Boeng Pes) and (vi) evergreen swamp forest. The water bodies are surprisingly scarce in Prey Long Forest, although it is drained with several small rivers, such as Siembok, Porong, O'Long, Chinit, O'Kachong and O'Ronoul and their smaller tributaries. There are also a small Lake Boeng Pes and some forest swamps locally called 'Cheum' (McDonald 2004). McDonald was the first who discovered and for the first time described the previously unknown peculiar evergreen swamp forest type of the Cambodian Lowland, examined by him (and now me) at the Cheum Takong forest swamp. He prefaced his description as follows: "Over the course of our survey we were able to confirm preliminary reports of extensive, discontinuous evergreen swamp forests in the region. Based on a thorough examination of botanical literature, I can only conclude that this unique vegetation type is exceedingly rare and endemic to the region. Moreover, it is unknown to science. " (McDonald 2004: 22).
I examined four sites in the Prey Long northern part as based at Spong village. The village is close to the left bank of a small rivulet of the same name, mostly hidden in vegetation, which serves a border between Preah Vihear Province (left bank) and Stung Treng Province (right bank, with the village). The village is surrounded by clearings of different age, small rice field and young cashew plantations, alternating with forest remnants. The rivulet is mostly hidden in impermeable vegetation as being quite small.
Cheum Takong forest swamp extends for some 3 km from SSW to NNE but only some 400 m wide, being a swamped valley of the O'Long River. It is occupied by a system of many small, shallow anastomosing courses of running water, often entering shallow pools with red, muddy, sucking bottom, forming a kind of 'inner delta' of the river. They abound in upright and pointed pneumatophores of some tree, which made the area strongly resembling mangroves. Cheum Takong is densely overgrown with bush and arboreal vegetation and so is very dark, with few sunlit spots. A large share in vegetation belongs to palms of several species, of which the most abundant is Licuala sp., some rattan (Calamus sp.) and the conspicuous, tall and elegant Livistona sp. At the SE end of Choum Takong there is a small (ca 150 m long) and narrow sunny open area (at which we camped), crossed by an old road and mostly occupied by a swamp with stagnant, turbid water filled with inundated fine grass. A very old, incipiently overgrown but yet sunny, forest road goes along the NW side of Cheum Takong. These grassy swamp, a small glade nearby and the road were sunny places where most butterflies occurred, which were almost absent from the Cheum Takong interiors.
I also briefly examined an already cleared area to the west of Prey Long Forest, mostly supplanted by young plantations, in Chey Saen District of Preah Vihear Province along Road 215, from Chey Saen village to slightly SE of Phneak Roleuk village. This area is crossed by several considerable rivers.

Stung Treng Province, the area west of the Mekong River (mostly after Kosterin 2017)
The region is most a flatland west of the Mekong River, with some sharp hills of tropical karst and a hilly ridge along the Mekong River at the magnificent waterfall of this river, known as Nimith (Nimet) or Labak Koun in Cambodia (at its left bank) and Khon Thai or Khone Pha Pheng (or, separately, Khone Falls and Pha Pheng Falls) in Laos (occupying the right bank and islands) (Fig. 1). This is the largest waterfall in South-East Asia, formed of a succession of rapids 9.7 km long and 10 km wide.
The land is mostly formed by soils of reddish-brown gravel and occupied by open low deciduous dipterocarp forest mostly formed by D. intricatus. This forest still covers most of the area of Thala Barivat and Srae Ruessei Communes, the eastern part of Sam'ang Commune and southern part of Preah Rumkel Commune; alternating with cashew plantations and sporadic rice fields predominating in the rest of the area. The stripe of several hundred meters along new excellent roads is devoid of old trees but abounds in saplings implying fast forest recovery. However, everywhere in this type of forest, numerous charred stumps and logs are scattered indicating regular low fires occurring. This forest is a very picturesque parkland landscape with sparse low but stout trees with very large dark foliage. From distance it seems that the ground is covered with a fresh-green lawn grass. In fact this is thickets of bamboo Vietnamosasa sp. ('prech' in Khmer) with very branchy, thin but woody stems and small narrow leaves, which is 1-3.5 m high and hard to penetrate. Some small areas have sandy soils and approach savannah-like veals (Khmer), with lower trees (mostly D. intricatus), short and sparse grass, quite a diversity of flowers (mostly gingers) and scattered cycads (Cycas revoluta Thunb.). These forests are crossed by many brooks flowing in corridors between 'walls' of the grassy bamboo, with mostly gravel and sometimes rocky beds and slightly turbid, opalescent water. Larger rivulets and rivers are usually hidden in thickets of giant thorny bamboo, also have gravel beds and slightly turbid water.
Smaller areas of tall evergreen or semievergreen forest, resembling that of the Cardamom Mts., are associated mostly with hill slopes. For instance, quite a large area at the road towards Preah Rumkel village is covered with tall evergreen forest; so is the small but sharp hill of Phnom Preahkonkha, with low deciduous forest occupying its northern foothill. As a rule, margins of any patch of evergreen forest is piled by felled and charred trees and is hard to enter.
The Mekong right bank is variable, mostly with a medium-high ground bluff and clayey bed, but also with rock outcrops and sandy patches. The current is considerable downstream of the Nimith Waterfall but soon the river becomes very calm and expands to the so-called Veun Nyang/Anlong Cheuteal pool inhabted by a subpopulation of Irrawady dolphins. In spite of this, there is an active project of Don Sahong Dam across Mekong to be constructed at the Nimeth/Hon Tai Waterfall.

List of localities
The locality data are provided in the geographical order, first of all by provinces. Provinces are denoted with two bold letters. To avoid confusion of numerals, localities are also denoted by conventional nicknames used in this series of papers. These nicknames follow the locality ordinal numbers after back slash, both underlined. To avoid confusion with the previous and forthcoming communications of this series, and since some localities can hopefully be revisited in future, the numeration continues that of the two previous communications (Kosterin 2019a, b). Disposition of the localities examined is shown in Fig. 1, based on Google Earth.
156\Srae Ruessei: a rivulet overgrown with prech (Vietnamosasa sp.) forming 'walls' about a men height, winding among a sparce low deciduous dipterocarp stand with the grassy layer formed by the same low bamboo or, on areas with more sandy soil, of sparcer and lower herbage of the 'veal' type, with many flowering gingers and presence of cycads, 9 km NNW of Stung Treng, 0-0.5 km NE of Srae Ruessei village,105. For more information about some of the localities examined, including the landscape photos, see my odonatological papers, as follows: for O'Som vicinity in 2015 see (Kosterin, 2015), for Stung Treng Province in 2016 see (Kosterin, 2017), for Phnom Tumpor Mt. in 2019 see (Kosterin, 2019d); more such papers devoted to other areas/years to be published.

Results
The list of butterfly species identified by photos is given below as accompanied with localities, dates and, where possible, sex. For some species a short comment is added. The taxa for the first time reported for Cambodia in scientific literature are marked with asterisk (*). These records, as well as some problematic ones, are illus-trated by photographs taken in nature. Rarity of taxa in the neighbouring Thailand is indicated according to Ek-Amnuay (2006) as §, § §, § § § from common to rare.

32-41) and Stung Treng
*Arhopala moolaiana maya (Evans, 1932) Corbet (1941), Evans (1957) and Seki et al. (1991) (which obviously 'inherited' many points from earlier to later ones), formal application of which would lead to failure in identifying these photos. But if to follow the option that the band is only partly' dislocated then these keys would lead to A. moolaiana. My photos (in which a bit of UPS is fortunately seen) (Figs 29-30) fit perfectly the specimens of this species and subspecies shown by Ek-Anmuay (2006) and Inayoshi (2020), in which UNH postdiscal band can also be considered as completely dislocated at vein 2.

Prosotas dubiosa indica
Prosotas sp. ST: 151\Cheum Takong (a grassy swamp), 5.12.2019 (Fig. 19); the same place, 6.12.2019 (Figs 16-18). Normally P. nora ardates has UNH spots in spaces 6 and 7 in echelon (Fig. 16), while these two individuals (photograpghed in the same place as a 'normal' one above) have them in line (Figs 16, 19). This character is not mentioned in literature as diagnostic in Prosotas but is often diagnostic elsewhere in Lycaenidae. Gerard Chartier (pers. comm.) has an opinion that this character is not diagnostic and these specimens are also P. nora ardates.  (Felder et Felder, 1860) of crawling erratically with many quick turns (Fig 44). I even suspected its having some developmental anomalies or even being a Cirrochroa × Algia hybrid. Its identification is based on a distinct and unjagged postdiscal lines but is still tentative as it does not fit too well to the species' general appearance. The female photographed on the same day at a different part of the same valley differed from C. tyche (see below) by smaller size and UPS more reddish, with a distinct, moderately jagged discal line and more expressed UPF dark borders.  Baray, 5.11.2018; ♂, 114\upstream of 2 nd bridge, 30.11.2019. ST: ♂♂, ♀, 151\Cheum Takong, 5-6.12.2019.
D. chrisippus is common in Cambodia and surely escaped recording until now because of lack of attention for that reason (but was mentioned verbally in the first communication of this series (Kosterin 2019a). M. ancon, A. agrata, A. alesia, A. allata, A. ammonides, A. atrax, A. licates, K. sindu, and P. ganga were estimated by Ek-Amnuay (2006) as rare in the neighbouring Thailand, A. auerlia, A. moolaiana, D. theda, D. bisaltide, A. ranga, E. recta, and B. oedipodea as uncommon and A. silhetana, P. aluta, C. surya, H. chromus, P. dan and B. cinnara as common.
Of localities where pristine evergreen forest is preserved, the richest one was the Kbal Spean vicinities (locs 105-107) at the eastern side of the western Plateau of Phnom Kulen Mts., where I found 39 species for two days of observations. Of localities with moderably disturbed natural vegetation, the richest place was in the vicinity of the Spong village (locs. 149-150) inside the vast Prey Long lowland rainforest, where I found the same number of 39 species for one day. There patchy clearead areas provide a lot of sunny forest margins which is a favourable place if not for butterfly habitation but at least for observation of them. Among secondary habitats affected by human activity for not less than a thousand years, the very narrow strips of open dry forest and long fallow lands on small glades between the walls of West Baray at its SE corner and the vast lotus swamps, in the western vicinity of Siem Reap City (locs 128-129), appeared an unexpectedly rich locality in spite of its poor appearance. Although missing rare species, at any season this place provides a great diversity of butterflies common in Indochina. In total I found as many as 42 species there. (This is in a curious contrast to the walls themselves, also bearing arboreal vegetation, and the the banks of the reservoir, which are fairly poor.) The periodically inundated banks of Lake Tonle Sap offer unique vast communities of temporarily flooded forest and scrub, where some unusual butterfly species could be expected. Although there are indeed species which seem to be more frequent there than elsewhere, e.g. P. columella and B. exclamationis, only two species, I. similis and P. aluta, were repeatedly found only in temporarily flooded scrub and are probably confined to it.
The Phnom Tumpor Mt. (locs 100-102) resides in the Phom Samkos Wildlife Sanctuary and so was earlier specially studied with respect to its butterfly fauna by Monastyrskii et al. (2011) in 2006 and 2010. Such species as P. corvus, K. inachus, T. diores, L. chandica and L. confusa, observed on the Phnom Tumpor were not found specifically on this mountain by Monastyrskii et al. (2011), although were reported for Phnom Kmoach Mt in the same sanctuary, while L. pardalis was not reported for the sanctuary as a whole.