SOUTH AUSTRALIAN MUSEUM ±±j/ j/ DESCRIPTION OF TWO NEW LARVAL SMARIDIDAE (ACARINA) FROM AUSTRALIA R. V. SOUTHCOTT Summary Two new Australian larvae of Smarididae (Acarnia) are described, Smaris arenicola sp. nov. and Sphaerotarsus monticolus sp. nov., both from South Australia, from soil and litter. Each is the second species known as larva in its genus; both earlier described species were also Australian. The term AME is introduced in the coding of the metric data of the larval dorsal scutum. DESCRIPTION OF TWO NEW LARVAL SMARIDIDAE (ACARINA) FROM AUSTRALIA R. V. SOUTHCOTT SOUTHCOTT, R. V. 1997. Description of two new larval Smarididae (Acarina) from Australia. Records of the South Australian Museum 30(1): 1-12. Two new Australian larvae of Smarididae (Acarina) are described, Smarts arenicola sp. nov. and Sphaerotarsus monticolus sp. nov., both from South Australia, from soil and litter. Each is the second species known as larva in its genus; both earlier described species were also Australian. The term AME is introduced in the coding of the metric data of the larval dorsal scutum. R. V. Southcott, Honorary Research Associate, South Australian Museum, North Terrace, Adelaide, South Australia 5000. Manuscript received 1 May 1996. The mite family Smarididae Vitzthum, 1929 is cosmopolitan, most members being known as adults. The adults (and deutonymphs) are small mites, of a generally flattened stature, living as predators in litter or under the bark of trees. These mites are mostly known from their adult instars. The larvae are probably all ectoparasitic on small arthropods, as is the case with most other members of the superfamily Erythraeoidea. Thus one Australian larva, Smarts prominens (Banks, 1916) , has been found as an ectoparasite of small Psocoptera (Womersley & Southcott 1941). A taxonomic and historical review of the family was given by Southcott (1961a). Seven genera are known as larvae (Southcott 1995). Of these two are Australian, each with a single described species; that of Smarts prominens (Banks, 1916) (by Womersley & Southcott 1941), and of Sphaerotarsus leptopilus Womersley & Southcott, 1941 (by Southcott 1960). The present paper describes a further new species in each of these genera, both from South Australia, and known only as larvae. Materials And Methods Specimens described in this paper were obtained by Berlese funnel extractions of samples of soil and litter in South Australia. They were collected as active larvae, and each was confined individually in a small tube and various small arthropods were introduced as potential hosts, in an effort to elucidate life histories. The larvae were eventually mounted in Hover's medium after first being cleared in 50% lactic acid solution. Microscopy was done with the use of Leitz Ortholux/Laborlux microscope equipped with phase contrast and polarizing facilities. Drawings were made with the use of a drawing apparatus. Chaetotaxic and other nomenclature and coding are as in Southcott (1961b, c, 1963, 1992, 1995). The code AME is introduced as a measure of the anterior central emargination of the dorsal scutum. (In my previous papers I had misinterpreted the use of this term as ASBM, which was introduced by Fain & Elsen 1987). AME may be formally defined as the distance between the anterior end of the scutum, in the midline, to the central anterior point of the scutum; hence AME=ASBa - ASBM (Fain & Elsen). The terms protorostral, deutorostral and tritorostral will be used for the anterior setae of the gnathosoma, following Newell (1957). All measurements are in micrometres (]im) unless otherwise stated. All specimens will be deposited in the South Australian Museum (SAM). Systematics Family SMARIDIDAE Vitzthum Smarididae Vitzthum, 1929: 69, 70; Womersley & Southcott 1941: 61; Southcott 1946: 173; 1948: 252; 1960: 149; 1961a: 174; 1961b: 413; 1961c: 133; 1963: 159; 1995: 57; Shiba 1976: 189. Erythraeidae Womersley, 1934: 218 (part). Smarisidae Grandjean, 1947: 1. For other synonymy see Southcott 1961b: 413— 414. R. V. SOUTHCOTT Genus Smarts Latreille Smarts Latreille, 1796: 180; Womcrsley & Southcott 1941: 63; Southcott 1946:175; 1948: 252; 1960: 155; 1961a: 176-177; 1961b: 438: 1961c: 133; 1963: 163; 1995:63. Sclerosmaris Grandjean, 1947: 3, 53. For other synonymy see Southcottl961b: 438- 439. Key To Larval Smaris — PDS 27-29 jam long. AL scutalae blunt- ended. Sternala I thickened, well setulose S. prominens (Banks, 1916) — PDS 28 45 |im long. AL scutalae pointed. Sternala I narrow, lightly setulose 5". arenicola sp. nov. Smaris arenicola sp. nov. (Figs 1A-D, 2A,B) Material examined Holotype. South Australia: Loveday, 7 8.ix.l985, R. V. Southcott, larva ACA2158 from soil and litter sample TX244, site Map: Department of Lands, South Australia , Barmera- Loveday, 2nd Edition 1:10,000 6929-26, MR 466055 (140°25'03 M E, 34°17 , 25"S), extracted by Berlese funnel live 6.x. 1985; to Hover's medium 28.1.1991. SAM. Paratypes. Same data as holotype, 21 larvae from same site and sample, ACA2150, 21 52— 2156, 2159-2173, extracted live similarly 24.ix.l985-10.x.l985; Hover's medium. SAM. Diagnosis of larva Anterior scutalae pointed. Gelli 65-72 um long. PDS 28^15 um long. Description of larva (from slide-mounted holotype, supplemented by paratypes) Colour in life red. Idiosoma 245 long, 180 wide; total length of animal 330. Dorsal scutum transverse, oval, with slight central anterior emargination, minutely porose. Scutalae slender, pigmented, AL pointed, PL blunt-ended, with adnatc to slightly outstanding setules. AL scutalae placed anterolaterally on scutum; PL placed posterolateral^; all near edge of scutum. Anterior sensillary setae placed about midway between levels of AL and PL scutalae bases; sockets of posterior sensillary setae lightly protruding, placed at about middle of posterior border of scutum. Anterior sensillary setae thin, but thicker than the filamentous posterior sensillary setae; all with slight setules in distal two-thirds, longer distally. Metric data as in Table I. Dorsal idiosomal setae about 41 in number, pigmented, blunt-ended, with short, barbed setules; anterior setae in indistinct rows across dorsum; posterior setae form a denser group. Ventral surface of idiosoma: two sternalae between coxae I, pigmented, pointed, lightly setulose; sternalae II vestigial, presenting as small domes c, 5 across, between coxae II; between coxae III two intercoxalae 28 long, similar to sternalae I. Behind coxae III c. 1 1 setae similar to dorsals, 27-36 long, arranged approximately 2, 4, 4, 1, with a denser grouping towards posterior pole. Legs fairly short and thick, similar to those of Smaris prominens (see Womersley & Southcott 1941); lengths (including coxae and claws): I 380, II 370, III 410. Coxalae and other leg scobalae pigmented, setulose, pointed. Supracoxala to leg I a slender, blunted peg, 5 long. Leg specialized setae: SoGeI.77pd(23), VsGeI.90pd(4), SoTiI.50d(38), SoTiI.70pd(33), CpTi.75d(8)+SoTiI.76d(28), SoTil.78d(12), VsTiI.83pd(6). SoGeII.51pd(15), SoTiIL03d(16), SoTiII.78d(10). SoGeIII.52d(14), SoTiIII.03d(12). Tarsus I with SoTaI.37d(45), curved, strong, pointed; FaTaI..40ad(6); tarsus I also with an accompanying trichobothriala to the solenoidala, with several long setules, at .44d, seta 40 long. Tarsus II with SoTaIL35d(18), slender, nearly recumbent. FaTall appears as a minute vestigial structure immediately alongside, i.e. posterior to, SoTalL All tarsi with a long, slender, blade-like empodium, 23 -27 long, over-reaching lateral claws, which are paired on each tarsus, about 18 long, each a pad with a number of long ventral onychotrichs. Gnathosoma robust, pyriform, 77 long by 75 wide (combined). Cheliceral digits S-shaped, 16 long, sharp-pointed. Protorostral (galeal) seta simple, slender, pointed, 14 long. Deutorostral seta simple, conical, 7 long. Tritorostral setae simple, slender, pointed, 21 long. Palpal setal formula 0, 0, 1, 1, 3, 7 (including solenoiodala and terminala (eupathidala)); all setae (except palptarsal solenoidala) pointed; femorala, genuala and tibialae 1 and 2 setulose. Odontus pointed, hook-like, with two dorsal supplementary spurs behind tip. Palpal supracoxala dorsal in position, a clavate rod 5 long. TWO NEW LARVAL SMARIDIDAE I. y 30 FIGURE 1. Smarts arenicola sp. nov., larva. A, B holotype; A, Dorsal view, on left legs omitted beyond trochanters; B, Dorsal idiosomal seta ('b' in A). C, D Paratypes: C, Teratological dorsal scutum of ACA2156; D, teratologics! dorsal scutum of AC A2 161. (Each to nearest scale.) R. V. SOUTHCOTT FIGURE 2. Smaris arenicola sp. nov., larva, holotype. A, Ventral view, on left legs omitted beyond trochanters. B, Tip of palp, ventral view, further enlarged. TWO NEW LARVAL SMARIDIDAE TABLE 1 . Metric data for Smarts arenicola sp. nov. larvae (*for maximum values) Holotype n range mean s.d. c.v. Character AW 47 22 43^19 45.5 1.37 3.0 PW 50 22 42-53 48.3 2.35 4.9 SBa 22 21 19-23 22.1 1.01 4.6 SBp 16 21 13-16 15.0 0.921 6.1 LX 8 21 6-12 8.81 1.21 13.7 ASBa 20 21 15-22 17.8 1.83 10.3 AME 2 21 1-3 2.10 0.436 20.8 1SD 15 21 14-17 15.0 0.805 5.4 L 38 21 34^39 35.8 1.34 3.7 W 63 21 58-72 63.7 3.44 5.4 AAS 16 21 14-16 15.1 0.831 5.5 A-P 20 21 15-20 16.9 1.64 9.7 AL 56 21 45-64 55.1 3.54 6.4 PL 43 22 36^7 42.1 2.83 6.7 ASE 40 21 36-44 39.7 1.93 4.9 PSE 65 19 60-73 66.2 3.92 5.9 DS 31-38 22 38^5* 41.0* 1.79* 4.4* 'Oc' 37 22 36^2 38.1 1.98 5.2 MDS 32 22 32-38 34.4 1.99 5.8 PDS 31-38 22 38-45* 40.8* 2.05* 5.0* Gel 68 22 60-69 64.7 1.93 3.0 Til 66 22 60-69 64.7 2.12 3.3 Tal(L) 68 22 64-73 68.5 2.39 3.5 Tal(II) 33 22 27-35 32.2 2.56 7.9 Til/Gel 0.97 22 0.92-1.07 1.00 0.0340 3.4 Gell 62 22 55-64 59.3 2.44 4.1 Till 66 22 62-68 65.6 1.89 2.9 TaII(L) 50 22 47-60 52.8 3.15 6.0 TaII(H) 27 22 20-31 26.4 3.06 11.6 Till/Gell 1.06 22 1.02-1.19 1.11 0.0490 4.4 Gelll 68 22 65-72 67.9 1.77 2.6 Tilll 93 22 86-93 89.9 2.04 2.3 Talll(L) 55 22 49-61 56.3 2.64 4.7 Talll(H) 26 22 20-29 25.0 2.29 9.2 Tilll/Gelll 1.37 22 1.28-1.38 1.32 0.0360 2.7 AW/ISD 3.13 21 2.75-3.36 3.04 0.187 6.1 ISD/A-P 0.75 21 0.75-1.07 0.887 0.0913 10.3 AW/A P 2.35 21 2.15-3.00 2.71 0.260 9.6 StI 18 22 18-29 22.4 2.97 13.3 Cxi 42 22 40-53 44.0 2.59 5.9 CxII 22 21 20-31 23.3 2.51 10.8 CxIII 31 21 26-36 32.1 2.78 8.6 Til/AW 1.40 22 1.32-1.56 1.42 0.0668 4.7 Tilll/AW 1.98 22 1.83-2.12 1.98 0.0666 3.4 AW/AL 0.84 21 0.70-1.02 0.829 0.05^9 7.2 AL/AAS 3.50 20 2.81-4.27 3.65 0.341 9.3 Tilll/Til 1.41 22 1.28-1.44 1.39 0.0402 2.9 Till/PW 1.32 22 1.22-1.57 1.36 0.0768 5.6 LAV 0.60 21 0.47-0.64 0.56 0.0416 7.4 PW/AW 1.06 22 0.93-1.18 1.06 0.0545 5.1 AL/PL 1.30 21 1.13-1.52 1.31 0.112 8.5 R. V. SOUTHCOTT Etymology The specific epithet arenicola means 'sand- dwelling' (Latin), a reference to the site of capture. Biology All larvae were captured free. In an attempt to find possible hosts larvae were introduced individually into small tubes, and various small arthropods introduced as potential hosts. These included various insects: small moths, small hemipterans (aphids, psyllids, a ?mirid), small Diptera and wingless Psocoptera; also a small spiderling. However, no attempt at paras itization by any of the mites was observed. Genus Sphaerotarsus Womersley Sphaerotarsus Womersley, 1936: 119; Womersley & Southcott 1941: 63, 73; Southcott 1960: 149; 1961a: 177; 1961b: 144; 1963: 211; 1995:57,63. Definition of larva One eye on each side. Anterior sensillary bases anterior to AL scutalae bases; posterior sensillary bases posterior to PL scutalae bases. Palpal tibial claw (odontus) apically simple or with narrow terminal split, and an outstanding basal accessory tooth. Genu I with 3-6 solenoidalae. Tibia I with four solenoidalae. Key To Larval Sphaerotarsus PDSto 127 umlong .. S. leptopilus Womersley & Southcott, 1941 PDS to 86 um long S. monticolus sp. nov. Sphaerotarsus monticolus sp. nov. (Figs. 3A,B, 4A,B) Material examined Holotype. South Australia. Mt. Lofty, 5 20.x. 1991, R. V. Southcott, larva ACA2556, in damp soil and litter in Eucalyptus obliqua forest, sample TX320, site at Map: Adelaide 6628 111, 1:50,000, MR 897282 (138 41'47" E., 34°58'03 M S.), sample collected 5.x. 1991, extracted live by Berlese funnel 20.x. 1991. Cleared with 50% lactic acid, mounted in Hover's medium. Paratypes. South Australia. Hope Forest, ll.ix.1980, RVS, in moss under Eucalyptus baxteri forest, at Map Reference Milang 1:63,360 617393 (138°36'19 M E., 35°18'06"S.), site TX198, extracted by Berlese funnel 1 4-23. ix. 1980. Stirling, at Map Reference Adelaide 1:50,000, 2nd Edtn., 934258 (138 o 44*0"E., 34°59 , 30"S.), ll.xi.1984, RVS, sample TX232 of soil, litter and grass under Eucalyptus hueberiana y extracted by Berlese funnel 17.xU984. Diagnosis of larva PDS to 66 u.m long. Description of larva (from slide-mounted holotype) Colour in life: idiosoma orange, legs pale orange. Idiosoma 325 long, 220 wide. Total length to tip of cheliceral digits 410; in life these dimensions were 300, 200, 380 respectively. Dorsal scutum trapezoidal, with rounded anterior end and acute posterior end; scutum minutely porosc. Scutum has a sclerotized ridge enclosing the anterior and posterior sensillary areas, of a tear-drop shape, the point posteriorad, and within the bases of the four scutalae. Scutalae pointed, with many fine, pointed setules. Anterior sensillary setae thicker than the filamentous posterior sensillary setae; all sensillary setae with fine setules in distal half of setal shaft. Metric data as in Table 2. Eyes 1+1, lateral to scutum, 18 across. Dorsum of idiosoma with 24 setae, pigmented, expanding slightly distally, blunt-ended, with many fine setules; setae arranged approximately 6,4,2,4,2,4,2. Ventral surface of idiosoma: sternalae I pointed, setulose; sternalae II absent; anterior to coxae III a pair of intercoxalae similar to sternalae I, 36 long, 60 apart. Behind coxae III are 10 pointed, setulose setae, arranged 4, 2, 2, 2. Legs somewhat longer than idiosoma; lengths (including coxae and claws): I 400, II 395,, III 495. Leg scobalae pointed, lightly setulose. Supracoxala to leg I blunted, 4 long. Leg specialized setae: TbGeI.44ad(25), TbGeI.44pd(22), SoGeI.60pd(23), SoGeI.67d(16), SoGeI.76pd(17), VsGeI.90d(3), TbTiI.48pd(28), TbTiI.53ad(29), TbTiI.54pv(27), TbTiI.54av(29), SoTiI.61pd(30), SoTiL67d(30), SoTiI.77pd(30), CpTiI.80ad(9)+ScTiI.82ad(30), SoTiI.86d(25), VsTiI.95pd(2). VsGeII.88d(l), SoTill. 1 ld(14), SoTilI.86d(10). SoTiIII.09d(18). Tarsus I with SoTaI.17ad(20), TbTaI.20pd(22), CpTaI.40ad(9)+SoTaI.42ad(22), FaTaL58ad(l), CpTaI.70d(6)+TbTaI.71d(29). Tarsus II with SoTaII.39d(20). Lateral tarsal claws with a few slender onychotrichs; empodium slender, over- reaching lateral claws. TWO NEW LARVAL SMAR1DIDAE B TF / ¥ 0.396) and the other two species (ratio < 0.395) and confirm previous findings (Kitchener et al. 1987). Vespadelus haverstocki and V. regulus cannot be distinguished using this ratio. Distribution and Habitat Figs 2 to 6 show the geographical distribution of each species. Initially these maps were drawn using only bats that were identified by electrophoresis or penis morphology. Bats identified on wing measurements were examined later and all fell within the distribution ranges already recorded for the species. Habitat information has been taken from Carpenter et al. (1978) Reardon and Flavel (1987) and Lumsden and Bennett (1995). TABLE 1. Means and ranges of forearm length (mm) of female Vespadelus species in South Australia. Species SD Range V. darlingtoni 35.80±0.99 (34.7-38.1) 9 V. finlaysoni 34.45±0.86 (33.0-35.5) 11 V. regulus 31.37±1.14 (29.3-33.1) 50 V. haverstocki 29.12±0.95 (27.0-30.9) 38 V, vulturnus 28.47±1.24 (26.2-30.6) 15 TABLE 2. Range of ratios of phalanx I : metacarpal of digit III of Vespadelus spp. of the ' regulus complex' in South Australia. Males and females included. Species Ratio V. regulus V, haverstocki V. vulturnus 0.34O-0.395 0.340-0.395 0.396-0.420 50 38 15 LITTLE BROWN BATS IN SOUTH AUSTRALIA 31 134° 26° 30° 34° 26° 30° FIGURE 2. Distribution of V. darlingtoni in South Australia. Vespadelus darlingtoni (Fig. 2) is confined to forested areas in the wetter southern and eastern regions of the State; Vespadelus finlaysoni (Fig. 3) occurs in arid northern areas of South Australia in mountain ranges and rocky country in association with mines and caves; FIGURE 4. Distribution of V. regulus in South Australia. Vespadelus regulus (Fig. 4) occurs in the south and east of South Australia in mallee woodlands and forests. It roosts in tree hollows and buildings. It is the only species found on Yorke and Eyre Peninsula; Vespadelus baverstocki (Fig. 5) is found over a wide area in the arid and semi-arid woodlands and shrublands of the State and roosts in tree hollows and buildings; 26° 30° -38° FIGURE 5. Distribution of V. baverstocki in South FIGURE 3 . Distribution of V. finlaysoni in South Australia. Australia. 32 L. F. QUEALE 26° 30° 34° FIGURE 6. Distribution of V. vulturnus in South Australia. Vespadelus vulturnus (Fig. 6) is found in woodlands and forests of the south-eastern region of the State, and along the River Murray. It roosts in tall trees along the River Murray and in the South East (Reardon & Flavel 1987). Vespadelus regulus and V. darlingtoni are the only species of the genus to date recorded on Kangaroo Island. There are several areas of sympatry for Vespadelus species in the State. The range of V. finlaysoni is shared with that of V. baverstocki (Figs 3 and 5) but the former species is not found in the regions occupied by the other two 'regulus complex' species. The northern limits for V. regulus and the southern ranges of V. baverstocki overlap in central Eyre Peninsula and the Murray Mallee (Fig. 4). In this region it has been suggested that V. baverstocki prefers Acacia spp. open woodland and V. regulus, the mallee (C. Kemper, T. Reardon pers. comm., Lumsden and Bennett 1995). Vespadelus vulturnus inhabits wetter, tall forested areas. Considerable overlap occurs between all three 'regulus complex' species in the River Murray region (Figs 4-6). Here a corridor of sympatry follows the 250 mm isohyet which coincides with the separation of eucaiypt (mallee) and Acacia spp. woodland (Griffin & MacCaskill 1986) and the well-watered riverine eucaiypt forest. Colour Differences in fur colour have been noted between Vespadelus species (McKean et al. 1978; Kitchener et al. 1987; Reardon & Flavel 1987). The fur of V. darlingtoni is very dark brown, as is the skin on the forearms. By contrast, the fur of V. finlaysoni is grey. The 'regulus complex' species share a third colouration type in which the fur is bicoloured, the hairs having brown bases and pale tips. The ventral surface of the body is lighter than the dorsal surface. Hence colour is not useful for distinguishing between the species of the 'regulus complex'. Electrophoresis To test the key (below) an additional sample of female bats, with accompanying frozen tissue, was identified using the key as 26 V. baverstocki, 10 V. regulus and four V. vulturnus. Then allozyme electrophoresis was performed on this sample. The allozyme comparisons confirmed that the morphological characters had correctly predicted the identifications of all 40 specimens. This confirmation indicates that despite overlap in some characters, the key can be used with confidence to distinguish the species. Species summaries Vespadelus darlingtoni may be easily distinguished from the 'regulus complex' species by its greater size (forearm length) and from V. finlaysoni by its dark colour and distribution. Vespadelus finlaysoni is similarly distinguished by its large size from the ^regulus complex' species. There is a size overlap with V regulus but there is no overlap in distribution. Vespadelus vulturnus is separated from the two larger species, V finlaysoni and V. darlingtoni on size r with no overlap in forearm measurements. A size overlap occurs rarely with V regulus and more commonly with V. baverstocki and may be distinguished from these on the basis of the phalanx I: metacarpal ratio. Vespadelus regulus may be distinguished from V. darlingtoni on size and colour, from V. finlaysoni on distribution and partly on size, and from V. vulturnus partly on size and wing ratios. Vespadelus baverstocki can be readily distinguished on size from the two largest species, V. finlaysoni and V. darlingtoni. It can be separated from V. vulturnus on the basis of phalanx I: metacarpal ratio. In many cases it can be separated from V. regulus on size. When this is not successful electrophoresis must be employed. LITTLE BROWN BATS IN SOUTH AUSTRALIA 33 Key To Female Vespadelus In South Australia 1. — Forearm greater than 33 mm 2 Forearm less than 33 mm 3 2, — Latitude south of 33°S V. darlingtoni Latitude north of 33°S V. finlaysoni 3 — Ratio of phalanx I: metacarpal of digit 111 greater than 0.39; latitude south of 33°S V. vultumus Ratio of phalanx I; metacarpal of digit ITT less than 0.39; locality variable 4 4. — Forearm usually greater than 30.5 mm; mallee habitat/coastal or latitude south of 34°S V. regulus Forearm usually less than 30.5 mm; arid habitat/inland or usually latitude north of 34°S V. baverstocki Note on males Although males can be identified by penis morphology, this can prove difficult in the field. The characters discussed above can be used to identify males. However my data indicate that male forearms tend to be shorter than females. TABLE 3. Ranges of forearm length (mm) of male Vespadelus species in South Australia. Species x SD Range n V. darlingtoni 34.9±0.51 (33.5-38.1) 14 V. finlaysoni 33.2±0.64 (32.4-34.4) 13 V. regulus 30.3±0.96 (28.5-33.1) 51 V. baverstocki 29.2±0.96 (26.7-30.7) 29 V. vultumus 27.7±1.33 (25.2-29.5) 16 Thus the key can be used with some discretion as an aid in identification of males. The one exception is that the identification key and species summaries do not distinguish easily between male V. regulus and V. baverstocki on forearm lengths (see Table 3). Acknowledgments I wish to thank M. Hutchinson for his advice and encouragement. I am also grateful to K. Bowshall-Hill for drawing the maps. I wish to thank the two referees whose comments greatly improved the final paper, and my husband B. McHenry for encouragement and help in completing the manuscript. References ADAMS, M., BAVERSTOCK, P. R., WATTS, C. H. S. & REARDON, T. B. 1987. Electrophoretic resolution of species boundaries in Australian Microchiroptera. L Eptesicus (Chiroptera: Vespertilionidae). Australian Journal of Biological Sciences 40: 143- 162. CARPENTER, S. M., McKEAN, J. L. & RICHARDS, G. C. 1978. Multivariate morphometric analysis of Eptesicus (Mammalia: Chiroptera) in Australia. Australian Journal of Zoology 26: 629-638. GRIFFIN, T. & MacCASKILL, M. (Eds) 1986. k Atlas of South Australia'. South Australian Government Printer: Adelaide. KITCHENER, D. J., JONES, B. & CAPUTI, N. 1987, Revision of Australian Eptesicus (Microchiroptera: Vespertilionidae). Records of the Western Australian Museum 13: 427-500. LUMSDEN, L. F. & BENNETT, A. F. 1995. Bats of a semi-arid environment in south-eastern Australia: biogeography, ecology and conservation. Wildlife Research 22: 217-240. McKEAN, J. L., RICHARDS, G. C. & PRICE, W. J. 1978. A taxonomic appraisal of Eptesicus (Chiroptera: Vespertilionidae) in Australia. Australian Journal of Zoology 26: 529-537. REARDON, T. B. & FLAVEL, S. 1987. 'A Guide to the Bats of South Australia'. South Australian Museum: Adelaide. 85pp. TIDEMANN, C. R., WOODSIDE, D. P., ADAMS, M. & BAVERSTOCK, P. R. 1981. Taxonomic separation of Eptesicus (Chiroptera: Vespertilionidae) in southeastern Australia by discriminant analysis and electrophoresis. Australian Journal of Zoology 29: 119-128. VOLLETH, M. & TIDEMANN, C. R. 1991. The origin of the Australian Vespertilionidae bats, as indicated by chromosomal studies. Zeitschrift fiir Saugetierkunde 56: 321 330. FOUR NEW SPECIES OF ANTIPORUS SHARP (COLEOPTERA, DYTISCIDAE) FROM AUSTRALIA, WITH NOTES ON A. FEMORALIS (BOH.) AND A. INTERROGATIONS (CLARK) C. H. S. Waits Summary Four new species of Antiporus are described from Australia: A. jenniferae sp. nov., A. hollingsworthi sp. nov., A. willyamsi sp. nov. and A. pembertoni sp. nov. The distribution of A. femoralis (Boh.) and A. interrogationis (Clark) in Australia is reviewed. A key to the Australian species is provided. FOUR NEW SPECIES OF ANTIPORUS SHARP (COLEOPTERA, DYTISCIDAE) FROM AUSTRALIA, WITH NOTES ON A. FEMORAHS (BOH.) AND A. INTERROGATIONS (CLARK) C. H. S. WATTS WATTS, C. H. S., 1997. Four new species of Antiporus Sharp (Coleoptera, Dytiscidae) from Australia, with notes on A. femoralis (Boh.) and A. interrogaiionis (Clark). Records of the South Australian Museum 30(1): 35-42. Four new species of Antiporus are described from Australia: A. jenniferae sp. nov., A. hollingsworthi sp. nov., A. willyamsi sp. nov. and A. pembertoni sp. nov. The distribution of A. femoralis (Boh.) and A. interrogaiionis (Clark) in Australia is reviewed. A key to the Australian species is provided. C. H. S. Watts, South Australian Museum, North Terrace, Adelaide, South Australia 5000. Manuscript received, 28 February 1996. The species of Antiporus were last revised by Watts (1978) who recognised six species. In 1984 Brancucci (1984) resurrected A. interrogaiionis (Clark) from synonymy with A. femoralis (Boh.). In this paper I take the opportunity to describe four additional species of Antiporus Sharp from Australia and to review the distribution of A. femoralis (Boh.) and A. interrogaiionis. One of the new species, A. jenniferae, is close to A. simplex Watts, one, A. willyamsi, has unusual sexual hairs on the front and middle trochanters and the other two belong to a group of dark coloured Antiporus with enlarged metafemora in the males (A. femoralis, A. interrogaiionis, A. wilsoni Watts, A. hollingsworthi n. sp. and A. pembertoni n.sp.). Males of this A. femoralis group are easily separated on the shapes of the metafemur, proclaw and aedeagus. Females are difficult to distinguish, other than A. interrogaiionis which can be recognised by the colour on the pronotum. The new species arc rare in collections, although this may in part reflect a lack of collecting particularly in south-west Australia. The collections from which specimens were examined are listed under the following abbreviations: AM Australian Museum, Sydney. ANIC Australian National Insect Collection, Canberra. BMNH Natural History Museum, London. MV Museum of Victoria, Melbourne. SAMA South Australian Museum, Adelaide. WAM Western Australian Museum, Perth. UQIC Entomology Department Queensland University, Brisbane. Systematics Key To Australian Species Of Antiporus Sharp — Protarsus with 2 claws Females 2 — Protarsus with 1 claw Males 10 : — Relatively large (> 6mm), predominantly reddish-yellow, elytra with 5-6 often incomplete long dark lines A. gilberti (Clark) — Not with above combinations of characters 3 — Protarsi black. Relatively small (<4mm) .. A. jenniferae sp. nov. — Protarsi reddish-yellow, same colour as rest of leg 4 At least the basal segment of mesotarsi enlarged in comparison with protarsi 5 — Segments of mesotarsi similar to those on protarsi or smaller 6 Elytron with small and large punctures, all mesotarsal segments larger than those of protarsi, elytra with lighter streaking in younger specimens ...A. willyamsi sp. nov — Elytra with uniform punctuation, only basal segment of mesotarsi enlarged, elytra uniformly dark red-brown to black . A. wilsoni Watts — Smaller (< 3.5mm) 7 — Larger (>3. 5 mm) 8 36 C.H.S. WATTS 7 — Elytron dark, usually with three reddish- — Protarsi reddish-brown, same colour as yellow spots, pronotum reddish-yellow rest of leg 15 with dark front and rear borders A. bakewelli (Clark) 15 — Elytron unicolorous, except for slightly . r i . paler sutural region and sides. Metacoxal — Elytron and pronotum uniformly red- ,. , A - , ■ , 01 , J r ,.,,,, lines evenly diverging anteriorly. Claw brown A. simplex Watts , ,J , , ■ /, -., £ on protarsus relatively straight, with 8 _ More or less uniformly dark reddish- small basal tooth (Fig. 21) .. brown, elytron tip often appears truncated A ' sim P lex Watts A. blakei (Clark) — Elytron with one or more pale patches. — Elytra normally black with or without Metacoxal lines much less divergent in lighter markings. Tip of elytron pointed . 9 antenor ^f than '" mid ^- Cla f on protarsus sickle-shaped, with large basal 9 — Black portion in centre of pronotum not tooth (Fig. 18) A. bakewelli (Clark) reaching front border A, interrogationis (Clark) 16 — A - femoralis complex. The species are best separated by characters of the — Black portion in centre of pronotum aedeagus and melafemur given in Figs reaching front border !_ 15 In addition the b i ack portion in .... A. hollingsworthi sp. nov. (W.A. only), thc centrc of fe pronotum does not reacn 1A. pembertoni sp. nov. (W.A. only, the front border in A interrogationis female unknown), A.femoralis (Boh.) whereas it does in the other species. 10 — Prominent cup-like tuft of long golden setae on mesotrochanters (Fig. 28). Segments of mesotarsus larger than Antip or us femoralis (Boheman, 1858) segments of protarsi ( Fl § s 7 ' 8 & 9 ) A. willyamsi sp. nov. ^ Number examined, 83. — Mesotrochanters without such a tuft of Brancucci (1984) clarified the relationships setae. Segments of mesotarsus smaller between this widespread and common species and than those of protarsus 11 A interrogationis (Clark). It can be separated r , , . ., from this species by characters of the aedeagus 11 — Metafemur produced on inside near apex , . e a u • *u a i * u r r T - and metafemur and in having the dark patch on " ** the middle part of the pronotum reaching the — Metafemur simple 13 anterior border. In this last character it is similar to the other members of the A. femoralis group. It 12 — Relatively large (>6 mm), predominantly differs from these in cha racters mentioned under reddish-yellow, elytron with several long ±c diffcrcnt spccics and in ust ratcd m thc figures. black lines. Metafemur with prominent narrow triangular process on inside near Distribution a P ex A - gUberti (Clark) Australian Capital Territory: Black — Predominantly dark-brown to black. Mountain, ANIC; Hackett, ANIC; Mt Coree, Metafemur broadly produced on inside ANIC; O'Connor, ANIC. New South Wales: near apex 16 Allyn River, ANIC; Galston, ANIC; Queanbeyan, ANIC; Sydney, MV, ANIC; Waratah, SAMA; 13 — Larger (length > 4.5 mm). Mesotibia Williams River, ANIC. South Australia: strongly triangularly produced in middle Bcachport, AM; Eyre Peninsula, SAMA; on outside A. blakei (Clark) Kangaroo Island, SAMA; Lucindale, SAMA; Port — Smaller (length < 3.7 mm). Mesotibia Lincoln, AM, SAMA. Tasmania: 11 km NNW simple or thickened and angular but Blackland, ANIC; Eaglehawk Neck, AM; never as above 14 Georgetown, MV; Hobart, MV; Launceston, MV; " ' Liffey Valley, ANIC; Waldheim, UQIC. Victoria: 14 Protarsi black in contrast to red-brown 4 km W. Ballan, MV; L. Barracoota, MV; Cape colour of rest of leg Otway, MV;10 km NW Dartmouth Dam, MV; La A. Jennifer ae sp. nov. Trobe, SAMA; L. Lcarmonth, ANIC;10 km NW FOUR NEW SPECIES OF ANT1PORUS 37 10 11 12 15 FIGURES 1-15. la, dorsal view of aedeagus of A. hollingsworthi; lb, lateral view of aedeagus of A. hollingsworlhi showing apical appendage; 2, proclaw of male A. hollingsworlhi; 3, postfemur of A. hollings- worthi. 4, dorsal view of aedeagus of A. interrogationis; 5, proclaw of male A. interrogationis; 6, postfemur of A. interrogationis. 7, dorsal view of aedeagus of A. femoralis; 8, proclaw of ma\t A. femoralis; 9, postfemur of A.femoralis. 10, dorsal view of aedeagus of^. pembertoni; 11, proclaw of male A, pembertoni; 12, postfemur of A. pembertoni. 13, dorsal view of aedeagus of A. wilsoni; 14, proclaw of male^. wilsoni; 15, postfemur of.4. wilsoni. 38 C. H. S. WATTS Lexton, MV; Montmorency, SAMA; Mt Buffalo, MV; Mudgee, ANIC; Myrtleford, MV; Oakleigh, MV; Reservoir, MV; Werribee, MV. Western Australia: 9 mi NE by E Esperance, ANIC; Geraldton, MV; King George Sound, ANIC; Mullewa, SAMA; Ravensthorpe, ANIC; Stirling National Park, ANIC; Swan River, SAMA; Wilga, ANIC. Antiporus interrogationis (Clark, 1862) (Figs 4, 5 & 6) Number examined, 62. As pointed out by Brancucci (1984), A. interrogationis is clearly distinct from A. femoralis with which I had previously synonymised it (Watts 1978). Apart from the aedeagus (Fig. 4) and metafemur (Fig. 6), A. interrogationis differs from the other members of the A. femoralis group by the dark patch on the middle part of the pronotum usually not reaching the anterior border. Distribution Australian Capital Territory: Black Mountain, ANIC; Paddy's River, ANIC. New South Wales: 37 km E Hay, SAMA. South Australia: Naracoorte, UQIC. Victoria: 4 km W Ballan, MV; Bright, SAMA, MV, AM; Buninyong, MV; Delatite, MV; Delley's Dell Grampians, SAMA; Dividing Range, SAMA; Eltham, MV; Glen Valley, MV; Hamilton, MV; L. Purrumbete, MV; L. Wendouree, MV; Linga, MV; Melbourne, ANIC, MV; Merrimans Creek, MV; Mitta Mitta River, MV; Montmorency, SAMA; Mt Buffalo, SAMA; Nelson, SAMA; Noorinbee North, MV; Six Mile Creek Dartmouth, MV; 8 km NE Toolangi, MV; Warburton, MV; Werribee, MV; Yarra River, MV; Yarrawonga, MV. Antiporus jenniferae sp. nov. (Figs. 25, 26 & 27) Description (number examined 63). Length 3. 4-3. 6mm. Oblong-oval, convex, widest behind middle, elytron somewhat elongate posteriorly, tip sharp. Reddish yellow, portions of elytra a bit darker, sutural lines narrowly darker and bordered with pale strip, apical portions lighter; protarsi darker almost black in some specimens. Strongly, densely, and evenly punctured throughout. Pronotum and elytron narrowly margined, elytron moderately serrate towards apex, except for short distance near tip. Apical segments of both maxillary and labial palpi weakly bifid at tips. Prothoracic process narrowly lanceolate, rounded, almost keeled in cross section, slightly narrowed between procoxae. Metacoxal lines raised, moderately separated, diverging slightly posteriorly, widening to about twice their narrowest width anteriorly. Some long setae arising from pit in centre of sternite numbers three and four. A line of long setae towards sides of elytron. Male: Pro- and mesotarsi moderately expanded, first and second segments as long as wide, claw on protarsus thickened, strongly bent near base, abruptly narrowing to sharp point at apex, with large basal tooth. Mesotibia robust, broadly but weakly indented on inner side in middle. Seta tufts on mesotrochantcrs somewhat thicker than on female. Female: Mesotarsi a little expanded, protarsi less so, Protarsus with two simple claws. Remarks A. jenniferae can be distinguished from all other Antiporus by its relatively small size, and its black protarsi. It appears closest to A. simplex and A. bakewelli. From A. simplex it differs by its black protarsi, elytral colour pattern which although less marked than in A. bakewelli is clearly present, the aedeagus having a broader tip and widening slightly in the middle compared with the more parallel-sided aedeagus in A. simplex and the spine at the base of the claw on the male protarsi being larger than in A. simplex. Based on the small numbers of specimens of both species available it appears to be a little larger. From A. bakewelli it differs in its black protarsi, less marked and more diffuse elytral colour pattern, the aedeagus with a broader tip, and the claw on the male protarsi being straightcr in comparison with the smoothly sickle-shaped claw of A. bakewelli. Distribution Most specimens are from Cape York Peninsula, but it is also known from Adelaide River, N.T. (in BMNH) and Synnot Creek in northern W.A. (in ANIC). Types Holotype male: "25K. N. Coen Qld. 29/9/84 C. Watts" in SAMA. Paratypes: 1, "12.40S 142°40E Qld Batavia Downs Hmsd. 17-23 Jun. 1992 T.A. Weir, at FOUR NEW SPECIES OF ANTIPORUS 39 A 16 * * 17 18 v~ I I 19 20 21 22 23 24 ■; y nh 25 26 27 FIGURES 16-27. 16 & 17, dorsal and lateral view of aedeagus of A. bakewelli; 18, lateral view of male^ hakewelli proclaw. 19 & 20, dorsal and lateral view of aedeagus of ,4. simplex; 21, lateral view of male A. simplex proclaw. 22 & 23, dorsal and lateral view of aedeagus of ,4. willyamsi; 24, lateral view of male A. willyamsi proclaw. 25 & 26, dorsal and lateral view of aedeagus of A. jenniferae; 27, lateral view of A. jenniferae proclaw. 40 C. H. S. WATTS light", in ANIC; 2, "12.40S 142.40E Batavia Downs Qld 03-10 Mar 1993 at light, I. Cunningham", in ANIC; 1, "16.318 126. IE CALM site 25/1 Synnot Ck W.A. 17-20 June 1988 T.A. Weir", in ANIC; 1 1 "25k N. Coen Qld 29/9/85 C. Watts", in SAMA; 1, "24M.S. Musgrave N.Q. 20.5.72 J.C. Brooks at light", in ANIC; 1, "13.34S 142.32 E Qld, 17 km N.W. by W. of Rokeby (Vardons Lagoon) 27 Oct. 1992, T. Weir, P. Zborowski, Small pool with debris", in ANIC; 56, "Qld Mt. Molloy 2 km S. 30/3/96, C. Watts", in SAMA. Antiporus pembertoni sp. nov. (Figs 10, 11 & 12) Description (number examined 1). Length 4.6mm. Oblong oval, dark red brown - black, front of head, sides of pronotum broadly and sides of elytron narrowly but diffusely light- testaceous. Ventral surface testaceous, sides of meso and metasterna darker, apical tips of metatibiae darker. Dorsal surface rather finely punctured, stronger laterally, punctures on head a little smaller than eye facets, punctures at rear of pronotum towards sides and on adjacent areas of elytra tending to join in roughly longitudinal lines. Ventral surface much more strongly punctured, strongly rugose-punctate laterally. Pronotum narrowly but clearly margined, elytron weakly so. Prothoracic process narrow, strongly raised, sides sub-parallel, bluntly tipped. Metacoxal lines well separated, sub-parallel in posterior third and anterior quarter, diverging in between so that anteriorly about twice the distance apart as posteriorly, area between coxal lines and forward onto mesosterum depressed. Male: Protarsi moderately expanded, protarsal claw large, slender, with well developed slender basal tooth. Mesotibia weakly indented on inner margin towards apex. Metafemur broadly expanded on inner edge at apex, some strong punctures and weak longitudinal ridging on expanded area. Female: Unknown Remarks More weakly punctured than the other species in the A. femoralis complex, with the large punctures along the front of the pronotum distinct and more than twice the size of others on the pronotum. Characters of the male aedeagus and metafemur are diagnostic. Known only from a single specimen I collected in a small stream in cleared land near Pemberton. Distribution Known only from unique type from near Pemberton, S.W. Australia. Type Holotype male: "W. Aust. 15 km N.W. Pemberton, 17 May 1987, C. Watts" in SAMA, Antiporus hollingsworthi sp. nov. (Figs 1,1a, 2, & 3) Description (number examined 15). Length 4.1^. 6mm. Oblong oval, convex. Dark red-brown, appendages lighter. Strongly evenly and densely punctate throughout, punctures smaller on head than elytra, punctures at rear of pronotum towards sides and on adjacent areas of elytra tending to join in linear longitudinal lines, punctures towards sides of metacoxae and sternites and on elytral epipleura very close together and tending to coalesce. Pronotum and elytra narrowly margined except for short distance at tips of elytra, lacking serration, sides of elytra before apex appear weakly and broadly flanged from some angles. Prothoracic process narrow, blunt, strongly carinate. Metacoxal lines sub- parallel in anterior and posterior third, rapidly diverging in middle. Male: Pro- and mesotarsi moderately expanded. First and second segments wider than long. Claw on protarsus strongly bent near base, straight with parallel sides in middle, rapidly narrowing at apex basal tooth subobsolete. Mesotibia a little thickened, weakly and broadly indented on inner edge in middle. Apical ventral edge of mesofemur expanded a little. Inner ventral edge of metafemur with large, broad, triangular expansion in apical third. Parameres robust, broad, aedeagus arrow- headed with thin protuberance above tip arising from neck of arrow. Female: Pro- and mesotarsi a little expanded; claws, meso- and metafemur and tibia simple. Remarks Antiporus hollingsworthi is a member of the A. femoralis group. It appears closest to A. wilsoni from the south-east of Australia which it resembles in the pronotal and elytral grooving, thickened male mesotibia and expanded metafemur. Antiporus hollingsworthi can be separated from both A. femoralis and A. interrogationis by the lack of pronounced paler areas on the head, pronotum and elytra, although some specimens of FOUR NEW SPECIES OF ANTIPORUS 41 A. hollingsworthi do have head, pronotum and elytral patches paler than the rest of the upper surface. The tendency of the punctures on the rear edge of the pronotum and the adjacent areas on the elytra to link up and form grooves is more pronounced than in either A. femoralis or A. interrogationis and the triangular expansion on the male metafemur is more rounded than in either of these species. A. femoralis and A. interrogationis also lack the thickened and indented mesotibia present in A hollingsworthi^ A. wilsoni and to a lesser extent in A. pembertoni. Antiporus hollingsworthi is more strongly punctured than A. pembertoni and the males have differently shaped metafemora (Fig. 3). It differs from A. wilsoni by characters of the aedeagus, by the indentation on the male mesotibia being more central, and in having a triangular rather than a rounded expansion to the male metafemur. Antiporus hollingsworthi differs from all other Antiporus species (and all Australian Dytiscidae) by the apical appendage on the aedeagus. Distribution So far A. hollingsworthi is known only from Perth and Carrington, W.A. and from two unlocalised localities in the South-west. However, of all the Australian regions this is the most poorly sampled for water beetles so it is likely to be more abundant than the few known specimens indicate. Types Holotype male: "W. Aust. Maidavale, 27th April 1990, C. Watts", in SAM A. Paratypes: 1 male, 7 females, same data as Holotype, in SAMA; 1 female "Perth 10/65 DE", in SAMA; 1 female "S.W.A. ?Rotnest, Edward", in SAMA; 1 male "JT201", in SAMA; 1 male, 1 female "R.P. McMillan, Carrington, 19.7.53", in WAM. Etymology Named after Rod Hollingsworth, a strong supporter of Natural History at the South Australian Museum. Antiporus willy amsi sp. no v. (Figs 22, 23, 24 & 28) Description (number examined 4). Length 3.4-3.6 mm. Oblong-oval. Widest behind middle. Testaceous, head, lateral quarters of pronotum, patches on elytra and appendages lighter, metafemur and tibia tend to be a little darker towards their apices, lighter patches on elytron tending to form broad longitudinal streaks. Strongly, densely and evenly punctate, punctures on discs of pronotum and head weaker, particularly on head, a few scattered much smaller punctures on disc of pronotum and elytra. Pronotum narrowly but distinctly margined, elytra less so. Prothoracic process with central keel, narrow between procoxae. Metacoxal lines raised, sub-parallel in posterior half and anterior quarter, diverging in between, about twice the distance apart anteriorly compared with posteriorly, reaching mesofemur. Some long setae arising from pits in centre of sternites three and four, and a line of long setae on dorsal surface of elytron close to side. Male: Protarsi expanded, second segment broadest, mesotarsi strongly expanded, first and second segments largest, sub-equal. Mesofemur and mesotibia stouter than profemur and protibia. Protarsus with single claw, strongly bent, long and narrow, with well developed basal tooth and weak bulbous area in middle. Protrochanters with well marked cup-like tuft of golden setae, mesotrochanters with a much larger scallop- shaped tuft of similar setae. Metafemur stout, ventral inner apical region with scattered large punctures and about four longitudinal grooves/ ridges. Female: Protarsus with two weak claws. Mesotarsi moderately expanded, first and second segments largest, sub-equal. Mesotibia stout. Remarks The species is easily recognised by the spectacular development of the setae on the trochanters of the male which is unlike anything I FIGURE 28. willyamsi. Meso- and metatrochanters of A, 42 C. H. S. WATTS have seen on other beetles. The very broad mesotarsi in the male and the striped elytral colour pattern in teneral individuals are also distinctive. Since collecting the original specimens in 1983, I have revisited the locality (a deep, narrow, spring-fed drainage ditch in acidic soil) a number of times without success. Biologically the area has now changed considerably from a healthy habitat with numerous species to one that is almost devoid of insect life although physically little changed. Distribution Known only from the type locality in the south- east of South Australia and from Healesville, Victoria. Types Holotype male: "10 km S. Robe, S.A., 1/83, C. Watts", in SAMA. Paratypes: 2 females, same data as holotype, in SAMA; 1 female, "Healesville, V. 12/68, C. Watts", in SAMA. Checklist Of Australian Antiporus Sharp A. bakewelli (Clark, 1862) A. hlakei (Clark, 1862) A.femoralis (Boheman, 1858) A. gilberti (Clark, 1862) A. hollingsworthi sp. nov. A. Jennifer ae sp. nov. A. pembertoni sp. nov. A. simplex Watts, 1978 A. willyamsi sp. nov. A. wiisoni Watts, 1978 A. interrogationis (Clark, 1862) Acknowledgments References The curators of the collections listed earlier are thanked for allowing me to examine specimens in their care. Mrs Vicki Wade typed the manuscript and Mr R. Gutteridge drew the illustrations. All these are thanked for their support and help. BRANCUCCI, M. 1984. Notes on some species of the genus Antiporus (Coleoptera: Dytiscidae). Aquatic- Insects 6: 149-152. WATTS, C. H. S. 1978. A revision of the Australian Dytiscidae (Coleoptera). Australian Journal of Zoology, Supplementary series No 57: 1-166. THYLACiSUS MEGiRL4.\f< A NEW SPECIES OF THYLACINE (MARSUPEAL1A: THYLACINIDAE) FROM THE ONGKVA LOCAL FAUNA OF CENTRAL AUSTRALIA PETER F. MURRAY MURRAY* P. F. 1997. Thyhwitws megirianr, a new species of thylacitie (Marsupialia: Thylaeinidae} from the Oneeva Locai Fauna oi' Central Australia. Records of the South Australian Atu.xeum 3G< I): 43-61 r Approximately 10% larger than a maximum-seized Tasmanian wolf, the laic M ioccne' early pliocene Thylaciniu megirinni n. sp, exhibits some similarities to the Sate Miocene thylacinid, T. potens Woodbwnc l9o7 arid early to mid-Miocene ^imiyaanus ificksom Muirhead and Archer 1590, while simultaneously expressing several derived features in common with the Recent ihylacLtiu. 7" cy^OLtphuius. Other ihan the presence of a tiny stylar cusp-Hlce strucuue distal to stylar cusp D on M\ which could be an artefact of wear, T. megiriani shows no defi title uniquely derived characters Lhat would exclude i\ from ancc-slry of the recently extinct T. cyn&cepttalus. Peter P. Murray, Museums and Art Galleries of the Northern Territory, PO Box 2109 Alice Springs, Northern Territory 0871, Austraha. Manuscript received 14 October, 1996, revised 3 Marth. 1 997. Fossil vertebrates were first described from the Alcoota Local Fauna al Alcoota (Fig. 1), northeast of Alice Springs, by Woodburne (1967). The strati graphically superposed Qngcva Local Fauna contains certain taxa lhat are more derived than those from the nearby Alcoota Local Fauna (Murray, Megirian & Wells 1993, Megirlon, Murray & Wells 19%). Among its characteristic species is a zygomaturine diprotodontid, Kolopsis yperus Murray* Mcgirian & Wells, which is probably synonymous with the Chcltenhamian- aged Zygoma f urns gtilt Stirlon 1967. A", yp&us is iiiore similar lo Zygomaturus species than is the Akoota Local Fauna species Koi&psis torus Woodburne 1967. Thylacinus megiriani appears to be analogously derived in comparison to the Alcoota Local Fauna thylacinc, T. patens. A direct ancestor of T. cynocephatus has not been identified, I! rosfroiis, T. spelaean and T. major are considered to be synonymous with T. cynocephatMS (Dawson J9S2). Other thylacinids are considered lo be too primitive or too specialised (Archer 1982, Muirhead & Archer 1990, Muirhead 1992, Wroc pcrs. com.). Ambiguity in the homology of the posterior stylar euspules on (he molars of T. cynocephcttus has resulted in a less than definitive resolution of late Tertiary thylacinid phylogcny. T. megiriani provides some additional: evidence pertaining to the homology of the cusps and some hypotheses of ihc phytogeny o\^ Ihe recent species of Thvlacinw are discussed. Matejuals And Methods As is often the case with Ongeva LF fossils, NTM P961S is not an ideal specimen, having been found in a number of pieces scattered throughout a hard nodule of calcareous sandstone. As fragments were extracted more or less individually, including fragments of the molar teeth, the specimen was glued together at ihe excavation silc along what appeared to be the likely contacts.. The fragile condition of the specimen dictated lhat several coals of Synocryl (Bcdacryl) be applied to harden and hold it together securely. However, this compound sometimes has ihe disadvantage of obscuring important morphological details, particularly when fine particles of the matrix become incorporated with the solution, sometimes suppressing surface detail and occasionally even mimicking small structures. So it was with this specimen, that certain features of the restoration seemed untenable, such as a slight overlap of thy P~- alveolus with that of the P 1 . As Thylacinus potcn.i has no appreciable diastema between these teeth, this arrangement appeared to be plausible, though further cleaning of the matrix and excess Synocryl from the specimen, in response to an observation by one of the referees, proved that it had not been restored corrcclly. This revelation provoked a more radical program of disassembly and cleaning of the specimen which I was initially afraid to attempt 135°E FIGURE 1 . Locality map; the Dngcva LF ih located un Akoota Station, near Alice Springs Northern Territory. due lo its extreme fragility. The new restoration has greatly improved the resolution of the specimen, to the extent thai many of my initial observations can be slated more confidently; though as will be discussed some features of the specimen remain equivocal. As the first draft of (his paper was based on the original restoration, I have briefly described and illustrated the modifications [ made lo the specimen to avoid confusion with, any errant copies of the earlier manuscript. 1 ani extremely grateful to my referees fur alerting me to (he possibility thai the specimen might be distorted, and for drawing to my attention several other morphological features, resulting, I believe, in a much improved account of the species. Systemaiics Family THYLACINIDAE Bonaparte 1838 Thyluiinus me&iritmin. sn, Hoiatype NTM P9618 (field ref. no, SQ107), left maxilla with P 1 , P\ M 1 " 1 . Type focatity Hill l h South Quarry, approximately 2 km southwest of the junction of Ongeva and Waite Creeks, Alcoota Station, Northern Territory, latitude 22 a 52'5, Longiludc, L34*2TE The Type THYLA ONUS MEWRIAM N .SP. 45 FIGURE 2. Photograph of the fcft maxilla of Thylaeinu* megiriani n. sp. (NTM P9G18); A, lateral aspect; B, palatal aspect. This photograph depidLs iht: field reiteration of the specimen in which the fragment containing P 1 and P 2 had been glued approximately S.O mm posterior 10 its Lorrea position, lcadiri|f to the initial impression rhat the posterior root of P ] was adjacent to, or sightly overlapped thai of P*. Note- also thai the M'istwistcd medially (H) and is nearly a centimetre out of alignment with MP. was recovered from a sandy siltstone channel deposit in the Watte Formation, approximately 10m above the fossitiferous lacustrine scdimenls containing the Alcoola Local Fauna (Murray, NU'iiiruin &. WdLs H'93. Megirian, Murray & Wells 1996), (Fig, I). Fauna Ongeva Local Fauna (Murray et at. 19931. Rock unit and age Waire Formation, Cheltenhamian equivalent, laic Miocene-?cailv Pliocene. Specific diagnosis A Eargc, long-snouted thylacinid, equal to, or larger than T- cynocephahs bul less massive than r. p&tens; much larger than N. dicks on? (Muirhead and Archer 1990) and T. macknesxi (Muirhead 3992). P 1-2 shorter and narrower than in T. potem, P in line with P 2 \ contrasting with T. poterts in which P 1 is obliquely oriented and situated medial to the midline of the canine alveolus, P 5 large as in T. patens. Premolar diaslcmata longer than in T. potens and in Ntmhacinus dicksoni? more comparable to T. cynocephalus. "M 1 J longer and narrower than in B FIGURE 3. Outlines depicting changes made to the initial restoration of the specimen; solid outline represents the new restoration; compare with Fig. 2. A, lateral aspect; B, palatal aspect. The specimen was taken apart at its primary joins, cleaned of excess Synocryl and glue, and reassembled. The anterior fragment of the rostrum was carefully cleaned revealing a labial expansion representing the anterior alveolus of P 3 . It was glued in place at what is probably the minimum distance from P 3 , as no definite contact could be identified. Removal of additional matrix from a fragment of palate that previously did not seem to fit anywhere on the specimen, revealed the inner lining of the anterior P 3 root and was glued in position. The M 3 was twisted out of alignment with the molar row by about 15°. This region of the specimen, which consisted primarily of impacted matrix and bone fragments was embedded in modelling clay, then softened in acetone and pushed gradually back into its correct position. The area above and internal to the M 3 was then reinforced with dental plaster. These changes resulted in a significant transformation of the occlusal outline of the specimen, which subsequently lost some of its resemblances to T. potens. THYLACINUS MEGIRIANI N.SP. 47 1 j ALV/C ALV/M4 B FIGURE 4, Drawing of corrected restoration of T. megiriani n. sp. (NTM P9618) left maxilla; A, lateral aspect; B, palatal aspect. Abbreviations: AC, precingulum; ALV/, alveolus; B, stylar cusp B; C, canine; D, stylar cusp D; E, stylar cusp E (brackets [E] indicate doubtful, or if worn, the approximate position of a structure); IEM, interdental embrasure; IOF, infraorbital foramen; MI. ..3, molars; ME, metacone; MJS, maxillojugal suture; PI ...3, premolars; PA, paracone; PC, postcingulum. T. potens though relatively shorter in proportion to their width than in T. cynocephalus. Agrees meristically with T. cynocephalus in that M 3 is slightly longer than M 2 , differing from T. potens and Nimbacinus dicksoni in which M 2 is slightly longer than M 3 . Length of molars exceeds anterior width from protocone to parastyle; stylar shelf considerably narrower than in T. potens; metastylar spur much larger than parastylar corner on M\ contrasting with T. potens in which they are more nearly equal-sized; stylar cusp D present on M 1- " 3 as in Nimbacinus dicksoni and probably as in T, potens; rudimentary stylar cusp E present on M 2 , absent in T. potens and in T. cynocephalus 48 P. F. MURRAY but present on M 1 2 in Nimbacinus dicksoni. Rudimentary anterior cingulum present on M 2 in contrast to Nimbacinus dicksoni in which it is well-developed on M'~\ T. potens in which it is present on M 2 ~ 3 and T. macknessi in which it is present on M 1 , the only upper molar known for that species. A small precingulum is present on M 2 of T. megiriani. Ectoflex of M 2 ~ 3 less pronounced than in T. potens but greater than in T. cynocephalus. Paraconal reduction on M'~ 3 comparable to that of T. cynocephalus; protocones on M 23 reduced, similar to T. cynocephalus. Infraorbital foramen opens above posterior half of M 2 as in T. cynocephalus, unlike T. potens in which it opens above M 1 ; maxillojugal suture nearly reaches the margin of infraorbital foramen, approaching the condition in T. cynocephalus and in contrast with Nimbacinus dicksoni in which it terminates considerably short of the margin of the foramen. Etymology It is my privilege to name this species after Dirk Megirian, Geologist, Museums and Art Galleries of the Northern Territory, Darwin, NT; who discovered and meticulously extracted the specimen. Description Maxilla Thylacinus megiriani n. sp. consists of a maxilla that was recovered from nodular calcareous sandstone matrix in about a dozen pieces (Figs 2-4, 7). All but a few small fragments have been assembled into the left side of the animal's snout which includes the base of the zygomatic arch to just below the orbit and the infraorbital foramen and canal, to the posterior margin of the canine alveolus. The full extent of the missing jugal is clearly inscribed by its suture. In palatal aspect the maxilla is shallowly concave on the labial side and the snout was constricted behind P ! much as in T. cynocephalus. The base of the zygomatic arch is convex and marked by a pair of low, irregular crests. The higher one, corresponding to the maxillojugal suture, extends obliquely upwards to immediately behind the opening of the infraorbital foramen. The infraorbital foramen is a large, approximately 9.0 mm by 10.5 mm opening situated 21.0 mm above the approximate middle of the posterior half of M 2 (Figs 4, 7). Its position in T. megiriani is similar to that of T. cynocephalus and differs from the more anteriorly situated opening in T. potens, above M 1 . Anterior to the infraorbital foramen, a shallow, oval fossa extends to above P 3 , anterior to which, the lateral margin of the rostrum becomes convex. A fragment of the maxillary palate extending from P 1 to P 2 is broken at the median suture, A prominent median palatal ridge is apparent, parallel to which runs a deep, 5,0 mm wide palatal groove. The palatal width at between the posterior roots of P 1 would have been about 24.0 mm, approximately 5.0 mm wider than in T. potens. The palate is narrower at this point in T, potens because the posterior roots of P 1 are angled inwards and are also larger, whereas they are aligned with the median base of the canine anteriorly and the P 2 posteriorly in T. megiriani and T. cynocephalus (Figs 2-5). Three additional fragments of the palate can be approximated to adjacent structures medial to P 3 - M 1 (Figs 3, 4) as additional removal of matrix and Synocryl revealed the inner margin of the P 3 alveolus on the largest of the fragments. The premolar diastemata are similar to those of T. potens in that the P 1 and the P 2 are widely separated, whereas the diastema between P 2 and P 3 is shorter (Figs 5, 6). In the original restoration of T. megiriani (Figs 2, 3) it appeared that the posterior root of P 2 might have overlapped the labial side of the anterior root of P 3 by about 3.0 mm. The specimen is broken at this point and the contacts were not quite congruent, causing the anterior part of the snout to twist out of alignment. This contact was taken apart and cleaned, revealing more of the P 2 alveolus and a straight section of diastema, indicating that a P 2 /P 3 diastema of at least 5.0 mm is actually present (Fig. 6). The premolars of T. megiriani are large and posteriorly directed, though in proportion to the molars, the First and second premolar crowns and their alveoli are absolutely and relatively smaller than those of T. potens (Figs 2-6, Table 1) but the P 3 crown, though not the alveolus, is similar in size to that of T. potens. The lengths of the premolar alveoli of T. megiriani are as follows: P 1 , 10.5 mm; P 2 , 15.0 mm; P 3 , 18.5 mm. Corresponding measurements of T. potens are: P ! , -15 mm;P 2 , 17.5 mm; P 3 , 20.5 mm (Fig. 6). Reconstruction The reconstructed snout of T. megiriani is compared with that of T. potens and T. cynocephalus in Fig. 5. This reconstruction, based on modifications of the original restoration of the specimen (Fig. 3) indicates that although it is comparable in size and robustness to T. potens, its shape is more similar to that of T. cynocephalus THYLACINUS MEGIRIANIN. SP. 49 10 20 30 40 50 l I I FIGURE 5. Reconstruction of the rostrum of T. megiriani n. sp. for comparison with those of T. potens and T, cynocephalus, drawn to scale. Comparative lengths of the premolar alveoli are indicated by brackets. T. megiriani n. sp. is approximately the same size as T. potens, both species being considerably larger and more robust than the Recent Tasmanian Wo\f,Thylacinus cynocephalus. T. megiriani differs from T. potens in having a wider, less constricted palate at the level of P 1 , in its alignment of the premolars, and in having smaller P'~ 2 alveoli and longer P^diastemata, all features shared with T, cynocephalus. Distinctive features of the rostrum and palate of T. potens include the large size and oblique emplacement of P 1 , distinct narrowing of palate at the level of P 2 , separate rather than confluent P' and canine alveoli, emplacement of the P' internal to the canine base, the incisive foramina are formed within a deep, oval pit, (apparently absent in T. megiriani) and the palatal fenestrae are small. 50 P. F. MURRAY 15 6.2 17.5 2.0 20 j — i i FIGURE 6. Diagram comparing proportional features of the cheek tooth rows of A, Thylacinus potens, B, T. megiriani n. sp. andC, Thylacinus cynocephalus, all drawn to the same length from anterior P l to the metastylar tip of M 3 , measurements in millimetres. Refer to scale and actual measurements given above, as brackets show relative rather than scaled dimensions of the alveoli and the length of the metastylar spur from ectoflexus to tip. The P 3 alveolus in?", potens, damaged on both sides, is a composite restoration; note the relatively and absolutely longer diastema between P 1 and P 2 in T. megiriani, suggesting that the P 2 /P 3 diastema was probably slightly longer than restored. The two pieces were glued at this point because no additional surface was available to provide a secure union. Observe that the diastemata in T. cynocephalus are approximately equal lengths. in being more drawn out in the premolar region and in the premolar rows lying essentially parallel to each other, rather than converging at the level of P 2 . As the restored P 2 /P 3 diastema is a minimum dimension, the snout of T. megihani may actually have been slightly longer than depicted. The distinctive median pit for the incisive foramina in T. potens appears to have been absent in T. megiriani. Note also that the P 1 alveolus is confluent with the canine alveolus in T. cynocephalus and T. megiriani, whereas in T. potens they are separated by a bony ridge. Longitudinal palatal crests define the lateral margins of the palatine fenestrae in T. potens. Similar crests in T. megiriani are situated further apart, indicating that the palatal fenestrae were probably wider. Premolars P 1 is a small two-rooted premolar closely resembling that of T. cynocephalus. The crown is 8.5 mm long and 4.5 mm wide. The anterior THYLACINUS MEGIRIANI N.SP. 51 B D FIGURE 7. Diagram comparing the position of the infraorbital foramen relative to interproximal MVM 2 and its relationship to the maxillojugal suture in A, Thylacinus cynocephalus; B, Thylacinus megiriani n. sp.; C, Nimbacinus dicksoni (NTM P907-3) and D, Thylacinus potens. In this series, the infraorbital foramen in Thylacinus potens is positioned more anteriorly than in the other species under consideration. A more anteriorly situated infraorbital foramen is a primitive feature in didelphoids, dasyurids and probably thylacinids. However, the infraorbital foramen in Nimbacinus is situated more posteriorly than in T, potens, an otherwise much more derived species. Note that the maxillojugal suture lies well posterior to the foramen in Nimbacinus, whereas it reaches or nearly reaches the posterior margin of the foramen in T. potens, as it does in T megiriani and T. cynocephalus. This relationship is a derived feature, suggesting that the forward position of the foramen in T. potens is a secondary, rather than a primary condition, indicative of a derived state. Abbreviations: IOF, infraorbital foramen; LJS, lacrimojugal suture; MJS, maxillojugal suture. 52 P. F. MURJRAY margin of the crown commences with an enamel bulge about 3.0 mm above the alveolus. The posterior profile is steeper, longer and slightly concave, terminating in an enamel thickening 1.0 mm above the alveolus (Figs 2-4, Table 1) P 2 is represented by its anterior root and posterior alveolus. Its alveolus is about 15.0 mm long, 2.5 mm shorter than that of T. potens. The posterior alveolus is slightly larger in diameter than the anterior one. P 3 is considerably the largest premolar. The length of the crown at the base of the enamel is about 16.5 mm. The circular, heavily worn posterobasal cuspule is 8.5 mm transversely, FIGURE 8. Photographs of the left M 1 " 3 ofThylacinus megiriani n. sp.; A, M 1 " 3 , taken before correction of the position of M 3 on the specimen; B, enlargement of M 3 showing stylar cusp D and possible aberrant stylar cusp E, which may actually be a peculiar wear artefact; C, paste-up showing the approximately correct alignment of the M 2 metastyle with the parastylar corner of M 3 . Because of its dislocation, the occlusal surface of M 3 is tilted more lingually than those of M 1-2 , revealing the entire buccal side of the crown and resulting in the misleading impression that the metacone is situated more posterolingually than it actually is. Note the faint, postcingulum- like structure on M\ the small precingulum on M 2 and absence of same on M 3 ; The remnant stylar cusp E on M 2 is indistinct in A, but is faintly visible in C. THYLACINVS MEGIRIANI N.SP. 53 while the width across the anterior enamelled part of the crown immediately above the posterobasal cuspule is 7.0 mm. The crown is heavily worn on the anterior and posterior surfaces where distinct wear facets result in a steep, slightly biconcave profile. Molars. M 1 is heavily worn (Figs 2, 4, 8A, 9C). Both the paracone and metacone have been ground flush with the stylar shelf. The remnant base of the paracone indicates that it was probably reduced relative to the metacone. A remnant V- shaped 'wall' of vertical facets above the premetacrista and centrocrista are faintly visible on the specimen. Stylar cusp D, situated immediately posterior to the labial cleft, is the highest relief on the worn occlusal surface. Stylar cusp E is not apparent. The protocone is round and slightly constricted near its base. A faint cingulum-Iike crest extends from a low bulge located about mid-way between the base of the protocone and the metastylar tip, though its exact form and extent is obscured by a layer of calcite. The length of the tooth along the stylar crest is 14.7 mm. Its transverse width from the protocone to the parastyle is 12.3 mm and its posterior width from protocone to metastyle is 16.0 mm. The M' of T. megiriani is similar, but relatively longer than the equivalent tooth in T. potens (Fig. 9B) differing from that of N, dicksoni (Muirhead & Archer 1990) in its reduced stylar shelf, reduced paracone and absence of an anterior cingulum (Fig. 9A) and from that of T, macknessi (Muirhead 1992) in its retention of a distinct stylar shelf remnant with a large stylar cusp D, in the reduction of the paracone, in the absence of an anterior cingulum, and in having a steep facet or 'wall' above the paracone and metacone bases (Fig. 9D). M 2 is considerably larger than M 1 and broader transversely, though slightly shorter across the stylar wings than M 3 (Figs 2, 4, 8A, C, 9C; Table 1). Its length from parastyle to metastyle is 16.8 mm. Its anterior transverse dimension from protocone to parastyle is 15.0 mm and its posterior transverse dimension from protocone to metastyle is 19.5 mm. The parastylar corner is small compared to the metastylar corner. The smaller, lower and narrower paracone is cleanly divided from the metacone by a mesial continuation of the labial cleft. Approximately 3.0 mm distal to the labial cleft and 2.0 mm labiad of the metacone is a prominent stylar cusp D. It is connected to the metacone by a short crest. Separated by a shallow labial groove from stylar cusp D is a second, much smaller enamel expansion situated about 3.0 mm anterior the metastyle, which appears to represent a remnant stylar cusp E. Although not prominent, the tiny cusp has a small wear facet and a labial bulge on the stylar shelf is evident. The anterotingual surface between the protocone and parastyle is distinctly convex and emarginated by a short, though strong precingulum. The protocone is smaller than on M 1 and is more U-shaped than rounded. Remnants of the pre-and post- protocristae extend from either side of the apex of the protocone to the bases of the paracone and metacone respectively. A small, but distinct lobe is situated immediately above the dentine-enamel junction, approximately midway between the metastyle and the protocone. In overall shape, the M 2 of T. megiriani is structurally intermediate between Nimbacinus dicksoni and T. potens, but differs from both in having a much reduced parastylar wing, reduced paracone and relatively reduced height of the stylar shelf. There is no indication of stylar cusp E on T. potens, but the cusp is well -developed, though small, in Nimbacinus dicksoni, M 3 is slightly longer and more slender across the metastyle and protocone than M 2 . The anterolingual and posterolingual emarginations are reduced and the precingulum is rudimentary (Figs 2, 4, 8B, C, 9C; Table 1). The parastylar spur is smaller in proportion to its metastylar component than in the previous molar, though the ectoflex is more pronounced on M 3 than on M 2 . The length of the crown from parastyle to metastyle is 17.4 mm. Its anterior transverse dimension is 16.1 mm and its posterior transverse dimension is 20.3 mm. Though heavily worn, it is apparent that the paracone was originally low, narrow and much smaller than the metacone, which in turn, is high, transversely elongated and anteroposteriorly narrow. As in Thylacinus cynocephalus, the preparacrista is short and the internal angle formed by the preparacrista and postmetacrista is more obtuse than in T. potens. Prominent pre- and postprotocristae meet at the apex of the narrow, V-shaped protocone. Both crests extend to near the cusps of the paracone and metacone accentuating a furrow-like occlusal fossa between them. Stylar cusp D is situated -3.0 mm posterior to the labial cleft in the same position relative to the posterolabial base of the metacone as in M 2 . It is connected to the metacone by a low crest. A possible rudiment of stylar cusp E is situated about 4.0 mm anterior to the metastyle. Though small, it is more obvious than 54 P. F. MURRAY C if D [ 12 3 4 5 _L_I l_l MM C CCR B PPC PRMC 12 3 4 5 ME I I I I LJ MM LLLL MM POPC PRMC POPC / PRMC CCR FIGURE 9. Comparison of M 1 " 3 of T. megirianin. sp. with thylacinid species: A, Nimbacinus dicksoni (NTM P907-3); B, Thylacinus potens; C, Thylacinus megiriani; D, Thylacinus macknessi (M 1 reversed for comparison); E, Thylacinus cynocephalus. Abbreviations: AC, precinguium; B, stylar cusp B; C, stylar cusp C; CCR, centrocrista;D, stylar cusp D; E, stylar cusp E;ef,ectoflexus;gv, natural orwear groove in stylar shelf; MCL, metaconule; ME, metacone; MS, tip of metastyle; PA, paracone; PC, postcingulum; PCL, paraconule; PMC, postmetacrista; POPC, postparacrista; PR, protocone; PRMC, premetacrista; PSC, crest connecting metacone and stylar cusp D. THYLACINVS MEGIRIANI N.SP. 55 TABLE 1. Measurements of cheek teeth of Thylacinus megiriani compared with those of T. potens and T. cynocephalus. T. megiriani T. potens T . potens T. cynocephalus NTMP9618 CPC6746 UCMP 66971 sample means (Woodbume 1967) L 8.5 — _ — 6.2 P 1 W 4.5 — — 3.3 L <12.0 12.4 — 8.3 P 2 W <5.0 5.5 — 3.8 L 16.5 16.0 — 10.6 P 3 W 8.5 8.8 — 5.0 L 14.7 12.0 — 10.9 M 1 W, 12.3 12.8 — 11.5 w 2 16.0 13.5 — 11.5 L 16.8 15.7 — 13.2 M 2 w i 15.0 13.9 — 10.0 w 2 19.5 17.5 — 15.0 L 17.4 15.2 14.6 15.1 M 3 w t 16.1 15.9 14.7 12.0 w 2 20.3 19.0 17.5 17.8 the structure in the equivalent position on M 2 . It consists of a nearly circular raised enamel ridge with an oval wear facet or pit in the centre (Fig. 8 B, C). After long deliberation of the structure under the microscope, I have concluded that it could represent a freakish wear pattern, but that it also indicates that thicker, more obdurate enamel is located where stylar cusp E would be anticipated. The enamel on the molars of this specimen is uniformly quite thick and it may be the case that some of these small structures, including the postcingulum-like crests on M 12 might be the result of an exceptionally active enamel organ. A -3.0 mm wide attrition furrow between stylar cusp D and the equivocal cusp E extends vertically some 4.0 mm up the labial side of the stylar crest. Meristic gradients of the molars are as follows: metastyle to parastyle length, M'M 2 >M 3 ; Paraconc, M l >M 2 >M 3 ; Metacone diameter, M ! ?M 2 >M 3 ; Metacone height, M l ?M 2 M 2 >M 3 ; Metastyle length, M'M 2 >M 3 ; Stylar cusp E, ? M 1 , weakly expressed on M 2 , ? M 3 ; ?Postcingulum length, M'>M 2 nil on M 3 . Discussion The similarity in size and robusticity of T. megiriani to T. potens contributed to some of the errors in the initial restoration and interpretation of the fragmentary specimen. The new restoration of the fossil indicates that T. megiriani was a long-snouted species, proportionally more similar to T. cynocephalus than to T. potens. The length of the diastema between P 2 and P 3 is a minimum estimation, as no definite contact between the two fragments of the snout could be identified. The posterior margin of the fragment containing P 2 shows a definite lateral expansion for the anterior root of P 3 , but the actual alveolar margin and internal lining of compact bone is absent (Fig. 3). Consequently, Thylacinus potens appears to have undergone proportional changes in the rostrum, possibly a forward adjustment of the zygomatic root to shift the line of action of the masseter muscle more anteriorly. Apparently related to differences in the proportions of the snout is the position of the infraorbital foramen in T. potens. In that species the foramen is situated more anteriorly than in Nimbacinus dicksoni, yet the maxillojugal suture terminates close to the margin of the foramen as in T. megiriani and T. 56 P. F. MURRAY cynocephalus (Fig. 7). In Nimbacinus, the maxillojugal suture terminates a considerable distance posterior to the foramen. This particular combination of characters suggests that the condition in T. potens is secondary, or specialised rather than primitive, having occurred in conjunction with telescoping of the base of the rostrum. Another specialisation in T. potens, now apparent, is the obliquity and rather large size of the P l alveolus, which is part of an overall proportional shift of the cheek tooth row in relation the rostrum. In Nimbacinus dicksoni, the P 1 is relatively small and lies in direct line with P 2 -3 as in T, megiriani and T. cynocephalus indicating that the condition is derived. However, the position of the P 3 relative to the canine in T, potens is like that of Nimbacinus dicksoni in lying medial to the canine and in being separated from the canine alveolus by a small bony crest. These latter features are probably symplesiomorphic, whereas the relationship of the P 3 to the canine alveoli in T. megiriani and T. cynocephalus is probably synapomorphic. Comparison of the molar morphology among T. potens, T. megiriani and Nimbacinus dicksoni shows a similar mixture of plesiomorphic and derived character states (Fig. 9A-E). In some aspects, such as molar shape and development of the stylar shelf, T. megiriani resembles Nimbacinus dicksoni more than it does T. potens. In T. potens the stylar shelf is high and wide, yet appears to show more reduction of the stylar cusps than T. megiriani and certainly more reduction than in Nimbacinus in which the stylar cusps D and E are well-developed on M 1 2 , though cusp E is absent on M 3 . T. potens shows a much greater development of the ectoflex on M 2 ~ 3 than either T. megiriani or Nimbacinus dicksoni. These appear to be derived characters of the dentition of T. potens, which however, also shows some plesiomorphic states relative to T. megiriani and T. cynocephalus in retaining a well-developed precingulum on M 2 ~\ in retaining large protocone bases and in the length of M 2 exceeding the length of M 3 . T. megiriani is derived relative to Nimbacinus dicksoni and T. potens in M 3 being slightly longer than M 2 , in having more slender and proportionally longer metastylar spurs, in having reduced protocones and more reduced paracones, in the loss of the precingulum on M 3 and in having a more obtuse angle between the preparacrista and the postmetacrista. These features of T. megiriani trend towards the more derived states found in the molars in T. cynocephalus. However, T. megiriani is plesiomorphic relative to T. potens in retaining a vestige of stylar cusp E on M 2 and if the structure on M 3 represents a stylar cusp, it is autapomorphic (?reversal, eg. Marshall et al. 1994) among all known thylacinids in that respect. The heavily worn condition of M' of Thylacinus megihani renders comparison with the unworn equivalent of T. macknessi (Muirhead 1992) somewhat less informative. Thylacinus macknessi possess a well-developed anterior cingulum, which is clearly absent in T. megiriani. In T. macknessi the pre-and postprotocristae do not ascend the paracone and metacone bases to form acute crests as they do in all other thylacines including the M 1 of T. megiriani. The stylar shelf is well-developed on the M 1 of T. megiriani and much reduced in T. macknessi. The molar is broken in a critical position in relation to the position of stylar cusp D. In T. megiriani, stylar cusp D is very large and lies immediately adjacent to the metacone. This suggests that the cusp may have been reduced in T. macknessi. This comparison agrees in ail aspects with Muirhead's (1992) observations that T. macknessi is more derived in the development of these features than T. cynocephalus, yet plesiomorphic in the retention of the anterior cingulum and autapomorphic in its extension from stylar cusp B to the base of the protocone. Stylar cusp homologies among thylacines were more ambiguous when T. potens was the only fossil form of any antiquity known (Archer 1982). The expression of the cusps in Nimbacinus dicksoni is now well-documented, and accords with other members of the family Thylacinidae (Muirhead & Archer 1990). in Nimbacinus dicksoni, stylar cusps B, D and E are present on M'~ 2 and stylar cusp D is present on M 3 whereas cusp E is absent from that tooth. In T. potens, stylar cusp D is definitely present on M 3 and almost certainly was present on M'~ 2 , though the crowns are damaged in the region that the cusps are usually situated. There is no sign of stylar cusp E on M 2 ~ 3 in that species. In T. megiriani, stylar cusp D is well-developed on M 1-3 whereas stylar cusp E is rudimentary on M 2 . 1 am disinclined to attribute any phylogenetic significance to the cusp-like structure in the position of stylar cusp E on the M 3 of T. megiriani, although its putative existence is duly noted should future finds serve to either verify or negate its presence. In T. cynocephalus, stylar cusps are absent on M 2-3 in conjunction with the reduction of the stylar shelf A small stylar cusp occupying the usual THYLACINUS MEGIR1ANI N.SP. 57 Ni. dicksoni T. macknessi L potens T, megiriani 1 \ — \ \ 2 phalus FIGURE 10. Character distribution in species ofThylacinidae. 1-16 are apomorphic, 17-23 are plesiomorphic: 1, loss of 'wall' below postparacrista and premetacrista; 2, transverse narrowing of molars and loss of sty lar cusp D on M 2 ~ 3 ; 3, widened, elevated stylar crest, suppression of stylar cusp E at least on M 2 ; 4, increased ectoflex; 5, lengthened premolar diastemata; 6, confluent P'-C diastema; 7, elongation of metastyle, reduced parastyle, obtuse angle between preparacrista and postmetacrista, M i longer than M 2 ; 8, reduced stylar shelf; 9, reduced protocone; 10, loss of precingulum and increased concavity of anterior margin of M 3 ; 1 1, reduced ectoflexus; 12, posterior position of infraorbital foramen above middle of M 2 ; 13, reduced paracone; 14, loss of precingulum on M 1 ; 15, forward extension of maxillojugal suture to posterior margin of infraorbital foramen; 16, large body size; 17, (plesiomorphic) absence of diastema between P 2 /P 3 ; (plesiomorphic) notched precingulum on M 2 -M 3 ; 19, (plesiomorphic) P 1 situated lingual to canine; 20, (plesiomorphic) retention of precingulum on M 1 ; (plesiomorphic) 21, large paracone; 22, (plesiomorphic) metaconid on M 3 ; 23 (plesiomorphic) small body size. topological position of stylar cusp E is present on M 1 . This cusp was initially identified by Archer (1982) as stylar cusp E. Subsequently, Muirhead and Archer (1990) identified the cusp as stylar cusp D that had become reduced and migrated posteriorly in conjunction with the elongation of the metastyle. The predisposition of the stylar cusps in T. megiriani supports the latter conclusion, in that if stylar cusp E was present, it was already extremely reduced relative to stylar cusp D. Moreover, there is no satisfactory explanation of the fate of stylar cusp D had cusp E been retained, as its amalgamation with the metacone, analogous to Sarcophilus (Crabb 1982) appears to have been highly unlikely (Stephen Wroe, pers. com.). PhyJogenetic position of T. megiriani The character evaluations employed here are based on those of Muirhead and Archer (1990). As the new species exhibits additional structures, 58 P. F. MURRAY the character polarities pertaining to those structures are based on the observations made in the discussion. An hypothesis of the phylogenesis of T. megiriani is drawn from the following postulations (Fig. 10): Thylacinus cynocephalus-T. megiriani synapomorphies: lengthened premolar diastemata, confluent canine-P' alveoli, posteriorly situated infraorbital foramen (in line with middle of posterior half of M 2 ), elongation and narrowing of the metastyle of M 3 (including M 3 longer than M 2 , reduced width of parastyle, obtuseness of angle of postmetacrista and premetacrista and reduction of height of stylar shelf), greatly reduced paracone particularly on M 2-3 , reduction of diameter of protocone base, loss of precingulum and concave profile of anterior surface of M 3 , reduction of ectoflexus on M 2 " 3 . Thylacinus potens-T. megiriani-T. cynocephalus synapomorphies; loss of precingulum on M 1 , forward extension of maxillojugal suture to margin of Infraorbital foramen, possibly large size. Thylacinus potens-Nimbacinus dicksoni symplesiomorphies; absence of diastema between P 2 and P\ presence of precingulum on M 2 ' 3 , P 1 situated lingual side of canine. Thylacinus macknessi-Nimbacinus dicksoni symplesiomorphies: retention of anterior cingulum on M 1 , large (unreduced) paracone, possibly small size, remnant metaconid on M 3 . Thylacinus megiriani-Nimb acinus dicksoni symplesiomorphy: expression of stylar cusp EonM 2 . Thylacinus megiriani apomorphies: possible expression of stylar cusp E on M 3 (reversal?), possible expression of anterobasal cuspule or small postcingulum on M' 2 (?neomorphic). Thylacinus potens apomorphies: extreme development of labial emargination of molars, stronger ectoflexus on M 2 ~\ forward position of infraorbital foramen, obliquity and large size of P 1 , enlarged trigonid and/or reduced talonid in lowers. Thylacinus macknessi apomorphies: extreme reduction of stylar shelf, absence of crests ascending lingual bases of paracone and metacone, postparacrista and premetacrista connect low in the crown basin consequently the lingual wall below the paracone and metacone is absent. Thylacinus cynocephalus apomorphies: extreme elongation of metastyle on all molars, extreme reduction of stylar shelf, loss of stylar cusp D on M 2 " 3 . Probably because there are so few species and relatively few characters due to the fragmentary nature of the fossils, the resulting phytogeny seems to be fairly straight-forward (Fig. 11). Among the species considered, Nimbacinus dicksoni is by far the least derived form. In contrast, T. macknessi, while retaining more plesiomorphic features than any of the remaining species, is simultaneously more specialised than T. cynocephalus. It is therefore the least likely candidate for ancestry of any of the three later species (T. potens, T. megiriani and T. cynocephalus). T. potens is also highly derived in its extreme thickening of the labial emargination of the molars, deep ectoflex, relatively narrow talonid and possibly in some degree of rostro- facial proportional sliding. This species shows subtle modification of the relative proportions of the snout that are probably secondary because the forward extent of the maxillojugal suture relative to the infraorbital foramen appears to be synapomorphous with T. megiriani and T. cynocephalus. In several respects, T. megiriani is the least derived of the large species in retaining a rudimentary stylar cusp E on M 2 , in retaining large stylar cusps D on M 1-3 and in retaining a general proportional similarity in its molars to those of Nimbacinus dicksoni. However, other characters of T. megiriani show closer resemblances to those of T. cynocephalus than observed in any of the other species. In general appearance, the differences between T. megiriani and T potens are quite subtle. Its large size and robustness, relatively small, broad molars, large P 3 , presence of stylar cusp D and deep maxilla all bear a close resemblance to T. potens, to the extent that when I first saw the specimen in the field, I thought it might represent an extreme individual variation of the latter species. This resemblance biased my initial restoration of the specimen, although in spite of this, its detailed differences considerably outweighed the introduced and genuine similarities. It is an important feature of the specimen that while the state of many of its characters show an unambiguous trend towards T. cynocephalus, it is probably further removed structurally from the modern species than it is from T. potens. A reasonable stage of evolution estimation for the time of separation of T. potens and the lineal ancestor of T megiriani would be in the later part of the mid-Miocene. THYLACINUS MEGIRJANI N.SP. 59 T. cynocephalus T. meairiani T. potens 8 7 T. mackness Ni. dicksoni FIGURE 1 1 . Subjective dendrogram depicting hypothesis of phylogenetic relationships among thylacinid species; 1, loss of stylarcusp E on M 3 , loss of high postcingulum M'~\ plesiomorphic states of large paracone and small size; 2, more posterior position of infraorbital foramen; 3, reduced sty lar shelf, loss of 'wall' above postparacrista, centrocrista and premetacrista, extended precingulum M 1 ; 4, loss of precingulum on M 1 , forward extension of maxillojugal suture to margin of infraorbital foramen, large size, reduction of stylar cusp E on ?all molars; 5, widened, elevated labial emargination of molars, increased ectoflex on M 2-3 , loss stylar cusp E on at least M 2 ; 6, increased length premolar diastemata, reduced stylar shelf, reduced parastyle and paracone, lengthened metastyle, reduced ectoflex, more posteriorly situated infraorbital foramen; 7, possible neomorphic faint, low postcingulum on M 1 " 2 , possible stylar cusp E on M 3 ; 8, increased reduction of stylar shelf, increased elongation of metastyle, stylar cusp D confined to M ! , reduced ectoflex. Paleobiology and biochronology The Ongeva Local Fauna shows a distinct faunal change from the underlying Alcoota LF. A zygomaturine, close to, if not synonymous with Zygomaturus gilli is the dominant large herbivore, apparently replacing Plaisiodon centralis which it resembles somewhat in size and general morphology. A species of the crocodile Quinkana is relatively common, whereas the dominant Alcoota crocodilian, Bam, is absent. Though the sample is too small to be definitive, it would appear that T. megiriani had replaced the Alcoota LF thylacinid species, T. potens by Ongeva times. A new small ratite bird, possibly a dromornithid, is also present in the Ongeva LF (Megirian, Murray & Wells 1996). Continuity with the Alcoota LF is indicated by the presence of Kolopsis torus Woodburne 1 967, Dorcopsoides sp., Dromornis sp. cf. D. stirtoni and Ilbandornis sp. The Ongeva K. torus appears to be little changed from the Alcoota population, suggesting a fairly short interval between the two local faunas. Because the depositional 60 P. F. MURRAY circumstances of the two faunas differ, it is not entirely clear as to whether the samples represent distinct communities within a basically stable environment or are a reflection of more general, possibly fairly rapid changes in the biotic attributes of the region. Thy/acinus patens is a highly specialised thylacinid showing as least as much, if not more, derivation in its molar morphology from Nimbacinus dicksoni than T. megiriani. The increased width of the stylar area and relative widening of the trigonids and narrowing of the talonids of the lower molars, combined with its robust skeletal attributes, indicate that T. potens was capable of shearing and crushing highly resistant tissues, and may have been somewhat hyena- or Osteoborus-Yike in its habits. T. megiriani shows an incipient trend towards the extreme camassialisation that distinguishes T. cynocephalus; consequently it may have been more actively predacious in its habits than T. potens. It is likely that the two species would have considerable overlap in their habitat preference and behaviour. Given their close similarities and large size, it seems improbable that they would have occurred sympatrically. Such large predatory species usually have a considerable range, and it is therefore a more plausible speculation that T. megiriani represents a successional replacement of T. potens rather than an addition to the previously existing fauna. and the small -si zed, extreme shearing modifications of T. macknessi. The addition of this specimen to the small existing collection of fossil thylacines seems to help clarify and simplify our current understanding of the phylogeny of the group. Thylacinus macknessi indicates that a highly derived shearing complex had evolved in the early to mid Miocene among thylacinids, but did not lead to the development of the Recent thylacine species, T. cynocephalus, which appears to have originated in the late Pliocene. It is likely therefore that extreme camassialisation in the larger thylacinid species took place in the late Cainozoic, perhaps in response to the radiation of macropodine species that commenced in the early to mid Pliocene (Flannery 1989). A tempting speculation, based on the circumstantial evidence for periodic drought in Alcoota times (Murray & Megirian 1992) is that there may have been an abundance of carrion over a sufficient duration to have selected for a large mammalian scavenging species. Following subsequent climatic amelioration in the terminal Miocene, the hyena- like T. potens was out-competed by the more predacious T. megiriani. The subsequent loss of Thylacinus potens as a possible scavenging predator may have left the somewhat less productive scavenging niche open, recruiting a smaller species; namely, the ancestor of the living Tasmanian Devil, Sarcophilus harrisi. Conclusions As in T. potens, the new species has a large P 3 , retains stylar cusp D and has transversely wide molars. T. megiriani differs from T. potens in many of the features that also serve to distinguish the latter from T. cynocephalus. These include reduced ectoflex of M 2 * 3 and stylar shelf on M 3 , increase in the length relative to width of M 3 , narrow protocones, reduction of P 1-2 relative to the molars, alignment of the premolar row and infraorbital foramen opening above posterior M 2 . T. megiriani shows no definite structural features that would preclude it from ancestry of T. cynocephalus and shows a number of synapomorphous features including reduced paracone and lengthened postmetacrista. T. megiriani is less specialised within the morphological extremes of the large size and powerful shearing/crushing dentition of T. potens Acknowledgments I am indebted to Jeannette Muirhead and Alex Baynes for their useful comments on the first draft of this paper. Their observations led to the correction of a number of important details relating to the original restoration of the specimen and several key morphological features of which I was unaware were generously and courteously pointed out to me. I thank Dirk Megirian for his outstanding contribution to all aspects of the excavation, interpretation and documentation of the Alcoota/Ongeva Locality and, of course, for giving me the opportunity to describe and name this interesting species after him. I am very grateful to Ian Archibald, the NT Museum's Preparator/Taxidermist who photographed and printed the bromides under difficult circumstances. I also thank Rod Wells and the Flinders University students who participate in the Alcoota excavations annually. The field work at Alcoota is partially supported by an NT Heritage Grant. THYLACINUS MEGIRIAN! N.SP. 61 References ARCHER, M. 1982. Review of Miocene thylacinids (Thylacinidae, Marsupialia), the phylogenctic position of the Thylacinidae and the problem of apriorisms in character analysis. Pp 445-76 In 'Carnivorous Marsupials'. Ed. M. Archer. Royal Zoological Society of New South Wales: Sydney. CRABB, P. L., 1982. Pleistocene dasyurids (Marsupialia) from southwestern NewSouth Wales. Pp. 512-16 h 'Carnivorous Marsupials'. Ed. M. Archer. Royal Zoological Society of New South Wales: Sydney. DAWSON, L. 1982. Taxonomic status of fossil thylacines (Thy/acinus, Thylacinidae, Marsupialia) from Late Quaternary deposits in eastern Australia. Pp. 527-36 In 'Carnivorous Marsupials'. Ed. M. Archer. Royal Zoological Society of New South Wales: Sydney. FLANNERY, T. 1989. Phylogeny of Macropodoidea; a study in convergence. Pp. 447-506 In Kangaroos, Wallabies and Rat Kangaroos. Ed. G. Grigg, P. Jarman and I. Hume. Surrey Beatty and Sons and the Royal Zoological Society of New South Wales, Sydney. MARSHALL, C, RAFF, E. & RAFF, R. 1994. Dollo's law and the death and resurrection of genes. Proceedings of the National Academy of Science, USA. 91:12283-12287. MEGIRIAN, D., MURRAY, P. & WELLS, R. 1996. The Late Miocene Ongeva Local Fauna of Central Australia. The Beagle, Records of the Museums and Art Galleries of the Northern Territory. 13: 9-38. MUIRHEAD, J. & ARCHER, M. 1990. Nimbacinus dicksoni, a plesiomorphic thylacine (Marsupialia:Thylacinidae) from Tertiary deposits of Queensland and the Northern Territory. Memoirs of the Queensland Museum 28(1): 203-221. MUIRHEAD, J. 1992. A specialised thylacinid, Thylacinus macknessi, (Marsupialia: Thylacinidae) from Miocene deposits of Riversleigh, Northwestern Queensland. Australian Mammalogy 15: 67-76. MURRAY, P. & MEGIRIAN, D. 1992. Continuity and contrast in middle and late Miocene vertebrate communities from the Northern Territory. The Beagle, Records of the Northern Territory Museum of Arts and Sciences 9(1): 89-110. MURRAY, P., MEGIRIAN, D. & WELLS, R. 1993. Kolopsis yperus sp. nov. (Zygomaturinae, Marsupialia) from the Ongeva Local Fauna: new evidence for the age of the Alcoota fossil beds of central Australia. The Beagle, Records of the Northern Territory Museum of Arts and Sciences 10(1): 155-172. STIRTON, R. A. 1967. New species of Zygomaturus and additional observations on Meniscolophus, Pliocene Palankarina Fauna, South Australia. Bureau of Mineral Resources, Geology and Geophysics, Australia, Bulletin 85: 129-150. WOODBURNE, M. O. 1967. The Alcoota Fauna, Central Australia: an integrated palaeontologicai and geological study. Bureau of Mineral Resources, Geology and Geophysics, Australia, Bulletin 87: 1- 187. SOUTH AUSTRALIAN MUSEUM VOLUME 30 PART 1 JULY 1997 ISSN 0376-2750 CONTENTS ARTICLES 1 R. V. SOUTIICOTT Description of two new 13 J. OKUNO A new shrimp of the genus Rhynchocinetes from the Great Australian Bight (Crustacea: Dccapoda: Rhynchocinetidac). 19 S. PICHELIN Pomphorhynchus heronensis sp. nov. (Acanthoccphala: Pomph Lutjanus carponotatus (Lutjanidac) from Heron Island, Australi; udae) trom 29 L. F. QUEALE Field Identification of female iittlc brown bats Vespadelus spp. (Chiroptera: Vcspertilionidae) in Souti 35 C. H. S. WATTS Four new species of Antiporus Sharp (Colcoptera, Dytiscidae) from Australia, with notes on A. femoralis (Boh.) and A. intcrrogationis (Clark). 43 P. F. MURRAY Thylacinus megiriani, a new species of thylacine (Marsupialia the Onseva Local Fauna of Central Australia. ylacinidae) trom Published by the South Australian Museum, North Terrace, Adelaide, South Australia 5000 SOUTH AUSTRALIAN MUSEUM \JtnE .IT MR ^fi JANUARY A NEW SPECIES OF MULTIPHASED CORBULIPORA MACGILLIVRAY, 1895 (BRYOZOA: CRIBRIOMORPHA) FROM SOUTHWESTERN AUSTRALIA. P. E. BOCK & P. L. COOK Summary A new Recent species of the cribriomorph genus Corbulipora, C. inopinata, is described from several localities in the western part of the Great Australian Bight. All species of Corbulipora, both Recent and fossil, are known to occur in several phases. The form of the subcolonies, and the characters of the component zooids may be very different in each phase, and subcolonies are capable of separate existence. C. inopinata occurs in an encrusting phase, in a flustriform, ovicellate phase, and a bilaminar rooted phase. All zooids have some form of cribriomorph frontal shield, unlike some of those of the closely related C. tubulifera. The shield has, however, different characters in each phase. A NEW SPECIES OF MULTIPHASED CORBULIPORA MACGILLIVRAY, 1895 (BRYOZOA: CRIBRIOMORPHA) FROM SOUTHWESTERN AUSTRALIA. P. E. BOCK & P. L.COOK BOCK, P. E. & COOK, P. L. 1998. A new species of multiphased Corhulipora MacGillivray, 1895 (Bryozoa: Cribriomorpha) from southwestern Australia. Records of the South Australian Museum 30(2): 63-68. A new Recent species of the cribriomorph genus Corbulipora, C. inopinata, is described from several localities in the western part of the Great Australian Bight. All species of Corhulipora, both Recent and fossil, are known to occur in several phases. The form of the subcolonies, and the characters of the component zooids may be very different in each phase, and subcolonies are capable of separate existence. C. inopinata occurs in an encrusting phase, in a flustnform, oviccllate phase, and a bilammar rooted phase. All zooids have some form of cribriomorph frontal shield, unlike some of those of the closely related C. tubulifera. The shield has, however, different characters in each phase. P. E. Bock, Royal Melbourne Institute of Technology, Latrobe Street, Melbourne, Victoria 3001, and P. L. Cook, Associate, Museum of Victoria, Swanston Street, Melbourne, Victoria, 3000. Manuscript received 2 April 1997. The occurrence of distinct phases of correlated colony growth form and zooid morphology in species of Corbulipora has been described in some detail by Bock and Cook (1994; in press). Essentially, the function of the phases appears to be the same in all species. The first, ancestrulate phase is minute, encrusts shells and shell fragments, and may even have an interstitial existence. It establishes the colony and results from the settlement and metamorphosis of a motile larva. The second, erect phase is cellularine or flustrine, and arises from stalk-like kenozooids and autozooids growing from the peripheral pore- chambers of the zooids of the primary phase. This phase develops ovicells and large interzooidal avicularia. The second phase gives rise in various ways to a small, bilaminar or frontally budded third phase, which is known, in Recent species, to be anchored by numerous rhizoids. As the ovicelled second phase tends to be thinly calcified, the delicate subcolonies may easily be destroyed, or at least detached from their origins. The bilaminar phase maintains the position of the colony and develops further ovicellate subcolonies, which may be able to alternate with it more than once (Bock & Cook in press). Collections of bryozoans, made by the R.V. Franklin, using an epibenthic sled, from localities in the western Great Australian Bight in July 1995, have produced a wealth of specimens and species. These include subcolonies of C. tubulifera and of an undescribed species of Corbulipora, which has several distinctive features. Materials And Methods Specimens are stored in the Collections of the Museum of Victoria (MOV) and South Australian Museum (SAM). Specimens for scanning electron microscopy were coated with gold. The Stations from which specimens of Corbulipora were obtained were as follows: C. inopinata sp. nov. GAB 118 34°59'S, 1 19°00'E, 85m. Young non- ancestrulate colonies encrusting worn adeonid and lunulite colonies of bryozoans. GAB 098 34°39'S, 122°26'E, 156m. 1 phase-2 subcolony with numerous avicularia. GAB 093 34°32'S, 122°58'E, 95m. 1 phase-1 subcolony and 1 phase-2 subcolony with ovicells. GAB 083 34°21'S, 124°08'E, 180m. 2 phase-1 subcolonies encrusting Turhtella shells. GAB 049 33°53*S, 125°22'E, 156m. 4 phase-2 subcolonies, developing into phase-3 at tips, one repeating phase-2. GAB 056 33°19'S, 125°43'E, 73m. 2 large phase- 64 P. E. BOCK & P. L. COOK 2 subcoionies, one with ovicells, developing from phase-3 subcoionies. GAB 045 33°25'S, H$*5WE, 143.5m. 3 small phase-3 subcoionies with stalks. GAB 020 33°20'S, 129°18'E, 157m. 9 small phase-2 subcoionies. 2 developing from phase-3 subcoionies. GAB 013 33°06*S, 130°00'E, 101m. 2 phase-2 subcoionies, one with ovicells, and one phase-3 with numerous rhizoids and one phase-2 subcolony. GAB 014 33°16'S, 130°00'E, 155m. 2 phase- 3 subcoionies developing into phase-2 subcoionies, plus 5 isolated stalks. C. tuhulifera (Hincks) GAB 019 33°22'S, 129°19'E, 301m. 7 phase-2 subcoionies, and 2 phase-3 subcoionies. Systematics Order Cheilostomatida Busk, 1852 Superfamily Cribrilinoidea Hincks, 1879 Family Cribnlinidae Hincks, 1879 Genus Corbulipora MacGillivray, 1895 Corbuiipora MacGillivray, 1895:60; Wass 1975:168; Bock & Cook, 1994:33; in press. Corbulipora inopinata sp.nov. (Figs 1-6) Material Examined: HOLOTYPE MOV, F80665, GAB Stn 049, subcolony including phase-2 developing into phase-3 and then repeating phase-2. f. FIGURES 1-2. Corbulipora inopinata sp.nov. 1, subcolony showing phase-2 autozooids and two ovicelled zooids at the proximal end (arrowed). Intermediate zooids leading to two astogenetic generations of phase-3 zooids at the growing tip (GAB Stn 049) X22; 2, phase-3 subcolony with marginal pore chambers and autozooids with occluded orifices (GAB Stn 020) X40. NEW SPECIES OF MULTIPHASED CORBULIPORA 65 FIGURES 3-4. Corbulipora inopinata sp.nov. 3, phase-2 zooids with ovicells and simple frontal spines (GAB Stn 049, see also Fig.l). X80; 4, phase-3 zooids with complex frontal shields (GAB Stn 049, see also Fig.l) X70. Paratypes: Rest of material L755, GAB Stn 049 part. including SAM Description Corbulipora with colony including three subcolony phases. Primary phase encrusting, ancestrula not seen; primary triad zooids with frontal shield of 16 simple spines. All zooids with a distinct, smooth gymnocyst, and communicating through small pore-chambers. Remaining primary-phase zooids with frontal shields with 16-18 spines, separated by 3-4 series of small lacunae and with two concentric series of raised pelmatidia. Calcified orifice with 3 spines, lateral pair often bifid. Second, flustrine phase inferred to arise from primary phase. Fronds developing from stalk kenozooids and elongated autozooids, often in quadriserial groups; series bifurcating to form fronds 10-12 zooids wide and 150 astogenetic generations long, bifurcating up to 4 times. Autozooids with distinct gymnocyst and 14-16 flattened spines overarching the frontal membrane. Brooding zooids with enlarged, curved, sometimes medially fused oral spines and large hyperstomial ovicell. Ovicell frontal with median suture and paired entooecial areas. Aviculana scattered, rare in presence of ovicells. Rostrum elongated, narrows not raised distally; slightly expanded and rounded terminally, mandible hinged on a delicate bar, orientated distally. Tips of fronds with 3^4 astogenetic generations of zooids with intermediate morphologies leading to the third, bilaminar phase subcolonies. These form small, expanded groups of zooids with shields similar to phase-1 subcolonies, with two concentric series of raised pelmatidia. Zooids have a small orifice, sometimes partially occluded by the fusion of enlarged oral spines. Numerous rhizoids develop from marginal pore- chambers, together with stalk kenozooids and elongated stalk autozooids with a small opesia and marginal spines, which are the origin of repeated subcolonies of phase-2. 66 P. E, BOCK & P. L. COOK JW^ ' /flf ' ||| -f ■ %'■ •-:•■':- & 1 e '^1 ■ ▼ 1^^^ | I m^ k ^«r i ^ ■.. ^*^?p c- ■' ■i>-.:?-:?^$ Hi Ho, *'-H- IIP ■-::■%.■■■ Ik If ^ FIGURES 5-6. Corbulipora inopinata sp.nov. 5, phase-2 autozooids, note enlarged pair of distal spines (GAB Stn 098) XI 00; 6, phase-2 autozooids and avicularium (GAB Stn 098) X80. Etymology' fnopinus L. - - unexpected: referring to the occurrence of a second Recent species of Corbulipora with multiphase growth. Remarks The three phases of C inopinata resemble those of C. tuhulifera closely, but differ in several respects. The most noticeable difference is in the occurrence of a spinous shield in the autozooids of the second, flustrine phase. The spines are flattened and generally do not fuse centrally, although the enlarged suboral spines of marginal zooids may be fused. In addition, the avicularia of C inopinata are much narrower, and are not raised distally, as are those in C. tuhulifera, and the mandibles are hinged on a delicate, but complete bar. The palate is much longer in proportion to the proximal gymnocyst than in C. tuhulifera. Other differences are small, but distinct and consistent. Although no ancestrula has been preserved, one of the small, phase- 1 colonies encrusting Turhtella from Stn GAB 083 possesses a post-ancestrular triad of very small zooids. These all have a shield of simple spines, very similar to the one post-ancestrular zooid of the triad in C, ornata (Bock & Cook, in press, fig. 1). Subsequent zooids have two series of pelmatidia, which do not develop until later in astogeny in C. tuhulifera. The orifices of this stage are also wider than those of C. tuhulifera. No stalk zooids have been preserved arising from the encrusting phase m C inopinata, but by analogy with C. tuhulifera, these are presumed to be identical with those which arise from phase-3 bilaminar subcolonies. These consist of one or two pairs of kenozooids 2.5mm long, often accompanied or succeeded by one or more pairs of elongated zooids 2mm long, each with a small opesia surrounded by 8-12 short, simple spines. Thus the stalks differ from those of C. tuhulifera in being more robust, with more zooidal elements, and in the occurrence of autozooids with an opesia, rather than kenozooids, early in astogeny. The zooids of intermediate morphology at the tips of the fiustnne phase-2 fronds attain a complete NEW SPECIES OF MULTIPHASED CORBULIPORA 67 shield of phase-3 autozooids within 4 astogenetic generations. First, the flattened spines fuse medially, leaving a row of small lacunae. A single series of pelmatidia develops on the spines of the next generation of zooids, and a paired series, with 3-4 lacunae, develops on the shields of subsequent generations. The small calcified orifice is complete within the next generation or two. Although the general appearance of the autozooids of the bilaminar phase is very like those of C. tubulifera, they have a smaller orifice. Although these are often partially occluded by the oral spines, they do not usually exhibit the series of changes resulting in complete closure of the orifice which occurs in C. tuhulifera (Wass 1975). In one specimen from GAB Stn 020, the spines are enlarged and nearly occlude the orifice (Fig. 2). One of the specimens from GAB Stn 013 illustrates the relationship and probable functions of phase-2 and phase-3 subcolonies particularly well. It consists of a small bilaminar phase-3 subcolony comprising approximately 60 autozooids. More than 50 rhizoids arise from the marginal pore-chambers, and on one face of the subcolony these are involved with and adhere to minute shell and bryozoan fragments. On the edge of the other face, one quadriserial stalk, 15mm TABLE 1. Comparative measurements in mm of Corbulipora inopinata sp.nov. and C. tubulifera (Hincks). Lz, lz length and width of autozooid; lo, width of orifice; Lov, lov, length and width of ovicell; Lav, lav, length and width of avicularium; Lp, length of avicularian palate. Character C. inopinata C. tub\ ulijera Phase 1 (including primary zooids in mms) Lz 0.41 -0.70 0.45- -0.74 lz 0.35-0.50 0.38- -0.50 lo 0.12-0.17 0.10- -0.13 Phase 2 Lz 0.98-1.30 0.78- -1.32 lz 0.26-0.29 0.25- -0.33 Lov 0.21-0.23 0.23- -0.27 lov 0.24-0.26 0.28- 0.33 Lav 0.60 - 0.75 0.65- -0.96 lav 0.15-0.17 0.22 0.26 Lp 0.43-0.51 0.31 0.35 Phase 3 Lz 0.50-0.75 0.54 0.87 lz 0.37-0.50 0.32 -0.54 lo 0.09-0.10 0.10 0.14 long, consisting of two generations of quadriserial kenozooids and 10-12 generations of elongated autozooids, gives rise to a phase-2 flustrine subcolony about 30mm long, with more than 40 astogenetic generations. It seems almost certain that the one face of the bilaminar phase was buried in the surface sediments, whereas the flustrine frond extended above the surface. This subcolony is quite small; two flustrine subcolonies from Stn 056 extend up to 70mm and include 10,000 to 12,000 zooids, many of which are ovicelled. These would protrude well above the surface of the bottom sediment, allowing the larvae to be released easily. The similarities in phase structure found in C. inopinata and C. tubulifera reinforce the inferences previously made by Bock and Cook (1994; in press) as to the nature and relationships of the known phases in C. ornata and C. suggerens, from the Victorian Tertiary. They also strongly suggest that the ovicellate phase of C, suggerens was a thinly calcified cellularine subcolony type, which has not been preserved as a fossil. C. inopinata occurs in the western Bight together with C. tubulifera, but from shallower waters. In Bass Strait, and from Tasmania, C. tubulifera is known from a wide range in depth from 40m to 800m. It appears to be at the limits of its range south east of Eucla, in deep water. Wass (1975) first described Corbulipora oriparma, which is the bilaminar phase of C. tubulifera, from three localities. The holotype was from northeast Tasmania, the other specimens were from the eastern part of the South Australian coast. Subsequently, Wass and Yoo (1983) recorded C. oriparma from 12 localities, extending from near Perth to southwest of Eucla. These are similar to many of the localities from which specimens have been recently examined, and which have produced C. inopinata. None of Wass and Yoo's (1983) localities was from a greater depth than 158 m, far shallower than the most westerly known locality for C. tubulifera. Only examination of the specimens recorded by Wass and Yoo (1983) can decide their identity. Acknowledgments We should like to thank Dr Yvonne Bone (University of Adelaide) and the Master and crew of CSIRO R.V. 'Franklin' for the opportunity for one of us (P.E.B.) to take part in the sampling programme in July 1995, which led to the discovery of the species described here. We are also grateful to Mr D. McDonald for his help in preparation of this paper. 68 P. E. BOCK & P. L. COOK References BOCK, P. E. & COOK, P. L. 1994. Occurrence of three phases of growth with taxonomically distinct zooid morphologies. Pp 33-36 in Hayward, P. J., Ryland, J. S. & Taylor, P. D. (eds): Biology and Paleobiology of Bryozoans. Olsen & Olscn: Fredensborg. BOCK, P. E. & COOK, P. L. in press. A revision of the genus Corbulipora MacGillivray (Bryozoa: Cribriomorpha). Memoirs of the Museum of Victoria. BUSK, G. 1852. Catalogue of Marine Polyzoa in the Collection of the British Museum. Part 1 Cheilostomata (Part).: viii, 1-54, vi. Trustees of the British Museum, London. HINCKS, T. 1879. On the classification of British Polyzoa. Annals and Magazine of Natural History (5)3:153-164. MACGILLIVRAY, P. H. 1895. A Monograph of the Tertiary Polyzoa of Victoria. Transactions of the Royal Society of Victoria, 4:1 166. WASS, R. E. 1975. A revision of the Bryozoan Genus Corbulipora MacGillivray. Proceedings of the Royal Society of Victoria 87:167-173. WASS, R. E. & YOO, J. J. 1983. Cheilostome Bryozoa from the Southern Australian Continental Shelf. Australian Journal of Marine and Freshwater Research 34: 303-354. THE ARCHAEOLOGY OF LAKE SYSTEMS IN THE MIDDLE COOPER BASIN, NORTH-EASTERN SOUTH AUSTRALIA. ELIZABETH WILLIAMS Summary This paper presents the final report of a study of the archaeology of a number of lake systems located in the middle section of the Cooper Basin and neighbouring areas near Innamincka, South Australia. The paper outlines the results of archaeological fieldwork carried out during 1987 in one area within the region and relates this to other research undertaken in 1986 and 1989 in other parts of the middle of the Basin. THE ARCHAEOLOGY OF LAKE SYSTEMS IN THE MIDDLE COOPER BASIN NORTH-EASTERN SOUTH AUSTRALIA. ELIZABETH WILLIAMS WILLIAMS, E. 1998. The archaeology of lake systems in the middle Cooper Basin, north- eastern South Australia. Records of the South Australian Museum 30(2): 69-91. This paper presents the final report of a study of the archaeology of a number of lake systems located in the middle section of the Cooper Basin and neighbouring areas near Innamincka, South Australia. The paper outlines the results of archaeological fieldwork carried out during 1987 in one area within the region and relates this to other research undertaken in 1986 and 1989 in other parts of the middle of the Basin. E. Williams, World Heritage Unit, Department of Environment, Sport and Territories, Canberra, Australia Capital Territory, 2601. Manuscript received 3 1 June, 1995. The lakes forming the main focus of this study are located within an extensive linear dunefield, the Strzelecki desert, which forms part of Australia's continental dunefield. Although archaeological research has been undertaken along the main channel of the Cooper in the middle Cooper basin, particularly the Innamincka area (Hughes & Lampcrt 1980, McBryde 1987), the archaeology of the region's series of lake systems had not been studied until this research programme. Given that at least a number of the lakes hold fresh water for some considerable time, and that this is a strikingly unusual feature of arid areas generally, it is very likely that the lakes would have been the major focus of occupation for at least the recent period and possibly for extended periods in the past. As well, it is possible that landforms associated with the lakes such as source-bordering dunes preserve information about Pleistocene climates and changing environmental conditions, and possibly also about Pleistocene human occupation. A study of the lakes therefore had the potential to contribute to our knowledge of not only the recent occupation of the region but also that of past settlement, and possibly, the wider question of how Australia's arid zone itself was occupied. In structuring the project, a preliminary analysis of the region revealed considerable variation of the nature and types of lakes. The study was designed to examine this variability in relation to human settlement. The study was designed as follows. I examined the archaeology of examples of the three major types of lake systems in the region. These are the semi-permanent lakes of the Coongie system, fed by the Cooper's North- West Branch; small isolated salt lakes marking the edge of the Cooper Basin, and a series of ephemeral lakes to the north of the Basin fed by local drainage and precipitation. The results of this work are presented below. In the latter sections of my discussion I compare this information with that obtained from sites located in a relatively waterless core of the dunefield located in the southern part of the Strzelecki desert. The concluding part of the paper outlines a number of speculative hypotheses about the occupation of the region and of Australia's arid zone in general. Work carried out in the Coongie lake system has previously been reported on in Williams (1988) and that undertaken in the Strzelecki dunefield summarised in Smith, Williams and Wasson (1991). The current study draws this research together to present a comprehensive view of the region's archaeology. General Environmental Features Of The Region, Including Those Important For Human Occupation The study area lies in one of the most arid parts of Australia. The region is covered by an extensive linear dunefield and receives the lowest rainfall reading of any area of the continent (125 mm per annum). Despite the aridity there are reliable water sources here in the form of the Cooper Creek channel, its overflow system - the North- West Branch, and associated lakes fed by the channel systems. The amount of water in this drainage complex does not reflect local rainfall patterns and predominantly derives from received precipitation falling many hundreds of kilometres 70 E. WILLIAMS to the north-east during the wet season in the upper parts of the Cooper catchment in north-west Queensland. Given the arid setting of the region, water was a valuable resource and one of the most important features for human occupation. This is confirmed by the distribution of archaeological material within the Coongie lake system - site size and stone artefact density increases in relation to the location of permanent and semi-permanent water sources (Williams 1988: 59). Another important feature for occupation was the availability of good quality stone suitable for artefact manufacture. Stone sources are extremely localised here and are either scarce or entirely absent within most of the linear dunefield itself. Some stone is available in the form of pebbles carpeting the surface of the gibber plains or 'stony deserts' located on the edges of the study area but this source only exists in the form of small nodules, not necessarily suitable for all types of artefact manufacture. The most extensive areas of stone are found in outcropping areas of hard rock within areas of dissected tablelands located on the margins of the study area. Although the availability of resources such as water and stone were crucial, the two resources do not always coincide geographically. In the area immediately surrounding Innamincka where are found the extensive deep waterholes of the main Cooper channel and the abundant good quality stone resources of the locality's mesas and dissected tablelands, the two resources are abundant and lie close together. In the Coongie lakes system located within the stone-free linear dunefield, although water resources are extensive, good quality stone lies at least 60 km away. Therefore one of the major features of the archaeology of the region relates to how people balanced their need for stone with that for water and how this varied from one area to another. Because this feature is responsible for much of the patterning of the archaeology of the Basin and neighbouring areas, it is the major focus for this study. A second and related focus concerns the antiquity of human occupation of the region. This topic is reported on in detail in Smith, Williams and Wasson (1991) and is also discussed here. The structure of the paper is as follows. Given that the environment places such strong constraints on human occupation in the region, a general overview of climatic and geographic data is presented first, including material on environmental chronologies. This is followed by a brief summary of historical accounts of how Aboriginal people were observed to be using the area at the time of European contact, in order to present a picture of the occupation of the area at least for the recent past. The paper then reports on archaeological sites investigated, including an analysis of stone artefacts from a selected sample of sites. I conclude with a brief discussion of some of the theoretical implications of my work. The Contemporary Environmental Srtting A number of the environmental parameters for the region have already been presented in Williams (1988). This section does not repeat this material and instead provides more detailed information on localities previously discussed and new data on areas not referred to in other publications. The main focus of the study area is a series of lake systems within the Strzelecki desert. On the east and west the Strzelecki is bounded by stony country, comprising gibber plains and areas of dissected tablelands. The latter consists of the low, but locally steep escarpments and small mesas of the geological features known as the Innamincka and Cordillo domes. The aridity of the area has been noted above. Precipitation in the region is unpredictable and erratic, with no systematic seasonal patterning. Evaporation is extremely high, at least one order of magnitude greater than precipitation (Reid and Gillenl988: 16). Three of the lake systems which form part of the study area are discussed in some detail here. They are the freshwater lakes of the Coongie system, a ring of salt lakes which mark the boundary of the Cooper drainage system, and another chain of lakes which lie to the north of the system and to the west of Cordillo Downs station. For convenience I refer to this latter group of lakes as the 'Cordillo lakes'. The Coongie system The Coongie system comprises a series of large, shallow, irregularly-shaped lakes lying within the linear dunefield. The lakes are fed by a tributary of the Cooper, the North- West Branch. The system has a complex hydrology and is described here in its constituent sections from south to north. Closest to the North-West Branch feeder channel are three lakes: Coongie, Marroocoolcannie and Maroocutchanie, which are fed directly from the North-West Branch. Two lakes lying further ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 71 T^ JUT Stony desert (Cordillo Dome) * Cordillo Downs if] \ Stony desert \ \ ■ \K l f\A !\v\\\\0*to Lake Koonoomoorinna ,} v \ ' L 1 ' ^h>\sVi' i* t ffy \y%TiIlie Lakes y\ \ \\\\\\ \M*> I wllUw \ Stony desert (Innamincka Dome) FIGURE 1. Regional map of north-east South Australia showing the sites mentioned in the text 72 E. WILLIAMS ^5 Salt Lakes ^"^ Source bordering dunes FIGURE 2. Map of the Coongie and Cordillo lake systems. ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 73 north, Toontoowaranie and Goyder, are fed by interconnecting creeks from overflow from the first three lakes. To the north again are a further three lakes: Marradibbadibba, Lady Blanche and Sir Richard, which fill from overflow from Lake Goyder. In this paper I refer to all of the lakes collectively as the 'Coongie system'. A further series of lake and channel systems also have some connection either to the Coongie system or the North-West Branch channel. It is noted that although they do hold water at certain times (as observed by the explorer, John McKinlay in 1862), the Coongie system, at least in the recent past, held the most extensive and most reliable sources of water. Little is currently known about their hydrology and for this reason the lakes do not form part of this study. The North-West Branch which joins the Cooper west of Innamincka diverts most of the flow of the main Cooper channel northwards. For the area west of Innamincka relatively little water thus flows down the main Cooper channel on to Lake Eyre (Mike Steele, Innamincka store, pers. comm.). The Coongie system therefore catches most of the water coming down the Cooper channel. Flow down the Cooper derives from two sources: local precipitation and received rainfall which falls as far as 500 km upstream in the channel country of western Queensland. This latter precipitation falls mostly during the summer monsoon and contributes the major portion of the Cooper's flow. In contrast to local rainfall its timing is relatively predictable. Water from the summer monsoon rains takes about three months to come down the channel, first reaching the Coongie about late autumn (Barry Saunders, Innamincka region, pers. comm.). Despite this level of predictability, the amount of water which actually comes down the channel and the exact time of year when it arrives markedly fluctuates in relation to variability in the timing and effectiveness of rainfall in the upper Cooper catchment. Nonetheless, there are predictable water sources in the form of deep, permanent waterholes which are often a couple of kilometres in length, located along the channels of the main drainage systems such as the Cooper and the North- West Branch. With the possible exception of Coongie, the lakes do not hold water on a permanent basis. Oral tradition amongst current residents of the Innamincka region suggests that Coongie Lake itself holds water for most of the year, although the nature and amount of water it holds is difficult to quantify because of a lack of historical records. The other lakes in the system hold water for months at a time but the extent and duration is highly variable and depends on a number of factors. These include the amount of water coining down the Cooper from rainfall in the northern catchment, local precipitation, the position of the lake relative to the North-West Branch and the season of the year (evaporation is highest during summer). Despite variability in the amount of water held in the Coongie system, the presence of often extensive supplies of fresh water in an otherwise extremely arid environment is extremely rare within Australia. The region consequently supports a very productive and diverse ecology, but one which varies in relation to water levels in the lakes and channels. The variability does not decrease overall productivity, and in fact is responsible for the high biodiversity of the region. The chain of lakes, with its complex sequence of rising and falling water levels is biologically highly productive. The lakes provide a diverse range of constantly changing habitats and have the potential to support the largest number and widest range of resources in the region. The major resources found here are: birds, especially waterbirds; fish, crayfish and mussels; dry-land and aquatic plants, and some mammal species. Biological studies of the system reveal this very high productivity (Reid and Gillcn 1988, Reid and Puckeridge 1990). Reid and Gillen for example recorded 161 separate bird species in the Coongie system and estimated that during the year under study (1986 - a wet year when all lakes in the system held water for much of the time) 20 000 waterbirds permanently occupied the lake chain. (1988: 184, 198). Because of the interconnectedness of the system, rising and falling levels in one lake affect other lakes in the chain. Given that most of the plant and animal species in the region are oriented to specific environmental niches, variation in water levels consequently provides a much wider range of habitats than would otherwise be available along a permanent waterhole where water levels remain relatively constant. For example, although water levels in one lake may begin to fall with evaporation, certain species of waterbirds whose habitat is that of more shallow water will migrate in to the lake while those whose habitat is that of deeper water will move to another lake where water levels are rising. The differential fluctuation of water levels means that the lakes have the potential to carry the largest biomass of plants and animals in the region 74 E. WILLIAMS because they are providing a wide range of ecological zones. The permanent waterholes on channels and rivers carry a relatively lower biomass because their environments are much less variable. Water levels fluctuate significantly less in these deep waterholes, supporting a smaller, but more predictable range of resources. Salt lakes or 'salinas ' These irregular-shaped lakes lie in a ring along the northern and western boundary of the Cooper drainage basin and mark the boundaries of that system. They are located in the midst of the linear dunefield and are deflation hollows cut to groundwater level. This type of feature has been termed a 'salina' (Macumber 1980). Their salinity derives from the fact that in the region the groundwater is saline. Although no hydrological studies of the salinas in this area have been undertaken, it appears that for the recent past at least, they are permanently saline. In contrast to the Coongie system they therefore support relatively few plant and animal resources. The Cordillo lakes To the north of the Coongie system, beyond the salt lakes, lie another series of irregularly-shaped lakes. As with the salinas, no formal studies of these have been undertaken and there are no records of how often they fill. Cartographic and field observations of the locality suggest that there are at least ten lakes in the system, connected by a series of channels. It appears that they fill from local run-off from the stony country of the Cordillo dome to the east. Their hydrology, unlike the Coongie system, therefore reflects local precipitation. The lakes were dry when fieldwork in the area was undertaken (1987), despite the years 1986 - 1987 having been much wetter than average in terms of local rainfall. This suggests that considerable local precipitation is required to fill them and they thus remain dry for very long periods of time. They are not salinas though, and small pools of fresh water were present in 1987 in a few places along the channels linking the lakes. The predominantly dry character of the lake system suggests that it supports relatively few plant and animal resources in comparison to the Coongie, but the presence of some areas of freshwater indicates that more resources are supported than the saline lakes. The region's current environment is not necessarily the same as today as in the past. Over time there have been significant changes in climate, hydrology and landforms, with consequent effect on the human occupation of the region. Past Environments - Dunefield And Floodplain Research Much of the information on past regional climates has been summarised and presented elsewhere (Williams 1988) and so here I briefly refer to this work and also discuss new data which has arisen as a result of the study. R. Wasson has developed an environmental chronology for the region, based on stratigraphic sequences derived from longitudinal dunes and floodplain sediments (for references see Williams 1988). A summary of his findings and a discussion of their implications for the hydrology of the lakes follows. Wasson's work indicates that before 22 000 years ago the Cooper had a higher velocity than today and was flooding out to a much greater extent. Although geomorphological studies of the Coongie lakes themselves have not been undertaken, Wasson's data suggest that if the Coongie system was in operation during this period it would have held considerably more water than today. As well, groundwater levels for the region as a whole would have been higher, indicating that the lake depressions which are now salinas may have held fresh water, Current work does not indicate whether local precipitation was also high at this time, so that it is not possible to model processes which may have been occurring in the Cordillo lakes. Around 22 000 years ago, the Cooper's velocity decreased, Hooding of the outer floodplain decreased and it is likely that the water table lowered as well. Consequently it is likely that what are now salinas would have changed from fresh water at this time. Lake levels in the Coongie system may have become lower. About 20 000 years ago conditions were windier and more arid, activating a period of dune building throughout the linear dunefield. The dune sediment was drawn from the deposits formed from the continuing, reduced flooding of the outer floodplain. These climatic conditions and the consequent dune building persisted until around 16 000 years ago. At about 1 2 500 years ago the flooding-out of the outer floodplain ceased, removing the sediment available for dune building. Dune building began again in the later Holocene, although on a smaller scale than in the late Pleistocene. Wasson (1986: 79) notes that this event is enigmatic, since in contrast to the late ARCHAEOLOGY OF THE MIDDLE COOPER BASrN 75 Pleistocene there seems to have been no climatic change occurring on a scale large enough to be able to mobilise sediments. He has speculated that it is possible, although not yet conclusively demonstrated, that the phenomenon could be related to more intensive human occupation of the region. If such intensive occupation took place, the erosion resulting from Aboriginal firing of vegetation to aid the procurement of food resources, together with clearing of trees and other vegetation to provide fuel for campfires and earth ovens, may have caused greater mobility in sediments. Past Environments - Studies Of Lakeshore Features The environmental data presented above is derived from a study of dunefield and floodplain sediments and presents a broad regional chronology. Is it possible to develop a chronology for the lake systems themselves? A formal study of the geomorphology of the lake systems was beyond the scope of this study so I developed a smaller project, more complementary to the archaeological work. This involved an examination of whether landform features such as source-bordering dunes, which have the potential to contain information on lake chronology, had developed on lakeshore margins. These transverse dunes, Twidale's 'leeside mounds' (1972; 85-86), are similar to the lunettes of semi-arid regions. They are formed when longshore drift transports debris to beaches or lee shores of lakes. This sediment is then locally redistributed by the wind before being trapped by vegetation close to the lake margin. The work was complementary to the archaeological survey in the sense that it was anticipated that source-bordering dunes were also likely to be a focus for human occupation. Systematic survey of these features would not only have the potential to reveal information about environmental features but also about human occupation, and the chronology of such occupation. The project comprised an examination of examples of source-bordering dune features for each of the lake systems discussed earlier. Aerial photographs were used to identify the features and distinguish them from the linear dunes of the continental dunefield. The transverse dunes are paler in colour and generally trend east-west in contrast to the north-south longitudinal dunes. Once they were identified on the aerial photographs I travelled to them on the ground and systematically surveyed all exposures for any archaeological material. Access to the features by vehicle was often difficult and in a number of cases access was by helicopter. Regarding the dating of the dunes it is noted that at the time the work was undertaken only radiocarbon dating was available to the study. Thus if organic material suitable for such dating was not present in-situ the dunes were unable to be directly dated. The results of the work are presented below. The Coongie system Information on a number of the source- bordering dunes of the Coongie system is outlined in detail in Williams (1988: 58-60). A summary of the work is presented here along with new data on other parts of the system. The analysis of aerial photographs, maps and the survey work itself revealed that while some features such as colour of the source-bordering dunes are similar across the system there is major variability in the number, size and location of the features. A common feature is that all dunes are pale in colour and have little or no carbonate formation within their profiles, and they contain intermittent scatters of artefacts on their surface - typologically the artefacts appear to be late Holocene in age. All of these factors suggest the dunes are recent features. When cores of the dunes were exposed by erosion and these sections were examined, no organic or archaeological material was found in-situ, with the exception of one site at Lake Marradibbadibba - Lake Goyder 1 . Details of this site are presented in Williams (1988: 60) where it is noted that it dates to the relatively recent past 810-130 BP (ANU 5424) thus confirming the recent age of the transverse features. There is considerable variation in dune shape and alignment across the system. Some lakes had a low and amorphous single transverse dune (such as Toontoowaranie); others had a higher and more distinct single dune (Coongie, Marroocoolcannie). Still others had multiple transverse dunes, one behind the other (five in the case of Lake Lady Blanche, yet none on the neighbouring Lake Sir Richard, and five or more in the floodplain between Coongie and Toontoowaranie). Location of dunes relative to lakes ranged from north, to north-east and east. Given that the study of the dunes was only a preliminary one, it has not been possible to 76 E. WILLIAMS determine what factors are responsible for the variation in dune features. It is likely that since there is major variation in the shape and size of the lakes themselves and how they fill relative to channel features and other lakes, and marked fluctuations in the amount of water coming into the system and the timing of when these events occur, it is probable that these factors are responsible for much of the variation. The morphology and colour of the dunes indicates that they are of late Holocenc rather than Pleistocene age. Two hypotheses could explain their recent age. The first proposes that the lake system was active in the Pleistocene, but since tt is still in operation today, receiving what can be major flows of water and often flooding out extensively, any Pleistocene shore features have either been swept away by previous high-water events or are well buried below flood deposits. In 1974 for example, the entire system flooded out - aerial and satellite photography reveals water extending for hundreds of square kilometres with only the tops of linear and transverse dunes visible. Krieg and Callen note that the Coongic system is one of net sediment accumulation (1990: 60) and it is thus possible that many older lake-shore features have been disturbed or buried by sedimentation. The second hypothesis proposes that the lake system may not have been in operation in the Pleistocene. Recent work by Callen and Bradford (1992) presents data showing that the current situation, in which most of the ilow down the Cooper channel west of Innamincka is diverted north up the North-West Branch, may be very sensitive to minor changes in topology resulting from such factors as tectonic activity or shifts in sedimentation or the location of dune features (transverse and longitudinal). Slight movements on faults or in sedimentary features of only 1-2 metres within a large, low angle fan formed by the Cooper immediately west of Innamincka may be sufficient to create major changes in channel direction, and thus whether the Cooper flow is diverted either along the Branch, along another channel such as the main Cooper channel or even southwards down Strzclecki creek. Given this sensitivity to change it is possible that during the Pleistocene the North-West Branch may not have been carrying the main Cooper flow and the lakes may not have functioned as they do today. Because a detailed study of the geomorphology of the Coongie system has not been undertaken it is not possible to resolve these competing hypotheses. It is nevertheless possible to conclude that whatever the reason, the lake-shore features currently exposed within the area do not provide a window onto the Pleistocene. The salt lakes Two of the larger salinas in the region, Coorie Coorie Tillie and Lake Deception, were chosen for analysis. Although the lakes lie some 100 km from each other, the form of their transverse dunes is virtually identical, in direct contrast to the variability of the Coongie system. As well, the transverse dunes themselves are very different to those of the Coongie. Each lake has a low, flat, pale-coloured transverse dune on its northern edge. These have been extensively gullied to expose older-looking sediments which are redder in colour and more compact with a pronounced carbonate horizon. It is likely that further analysis would confirm that the sediments are of Pleistocene age. A systematic survey of the length of each dune, including an examination of the many sections of dune core exposed by erosion, revealed that the small amount of archaeological material present all lay within recent sediments on the dune surface, rather than in the Pleistocene core. In contrast to the Coongie, there are no large amounts of water coming into the system to modify or distort landform features. The salinas are groundwater features, and do not fill from any channels. The dunes' consistency in nature and form could result from similarity in formation and hydrology operating on the lakes. Unlike the Coongie, the transverse dunes appear to have the potential to preserve Pleistocene sediments but no stratified archaeological material is currently visible within their profiles. The Cordillo lakes This system has a set of source-bordering dunes different from both the Coongie system and the salinas. Like the salinas and unlike the Coongie system, there appeared to be a strong consistency in the number, form and structure of the features across the lake margins. On the northern edge of the lakes is a sequence of three transverse dunes trending east-west, and located one behind the other. The dune closest to the lake edge is quite low, less than a metre in height and pale in colour. Behind this is another, slightly higher, pale-coloured dune which is beginning to blow out into small longitudinal dunes. Behind this again is another higher dune, distinctly orange in colour with a pronounced carbonate horizon, which has almost completely blown out into longitudinal dunes. ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 77 The outermost dune is heavily eroded and it is likely that it is Pleistocene in age. Samples of carbonate from within the profile of the dune from Lake Koonoomorinna were dated using radiocarbon techniques and returned the relatively recent dates of 3930-80 BP (ANU 6182) and 4560-80 BP (ANU 6182). John Head of ANU Radiocarbon Laboratory notes that the samples are not dating the formation of the dune itself, only the last major phase of mobilisation of carbonates within the dune. Such mobilisation would have taken place during the last period of stable high- water levels in the region and the samples are probably dating this event (pcrs. comm.). A complete survey of all of the dunes from one lake, Koonoomoorinna, revealed that stone artefacts are scattered across the surface of each dune, but that no artefacts or organic material were present in-situ within the cores of the features, including the Pleistocene dune. An analysis of a number of samples of the artefacts is presented later in the paper but it is noted here that those seen had features characteristic of artefacts dating to the mid-late Holocene. It is suggested that the regularity in source- bordering dune spacing and form for the Cordillo lake system is, like the salinas, reflecting consistency in processes involved in formation and filling. The lakes fill from run-off within the local catchment, and in contrast to the Coongie system there is significantly more regularity in their hydrological processes and many fewer flood events to affect the morphology of lake-shore features. Like the salinas and in contrast to the Coongie system, the transverse dunes appear to have the potential to preserve Pleistocene sediments but, as with the salinas, no in-situ archaeological material is currently visible within their profiles. It is not known at this stage why the lakes preserve a group of three transverse dunes. Each dune may reflect a period of relative stability in lake levels so that the position of the dunes, one behind the other, may indicate a shrinking lake basin since higher levels in the later Pleistocene. Alternatively, the position of the dunes may reflect the fact that all lakes in the region, including the Coongie system, are slowly migrating southwards over time (Bob Wasson pers. comm.). The presence of the dunes may reflect both factors. Following this presentation of climatic and environmental information, the next section places the region in a more recent context by summarising historical accounts of how Aboriginal people occupied the area at the time of contact with European people. The section draws on information presented in Williams (1988) and Layton, Foley and Williams (1991: 258-259) and also presents new material. Historical Accounts Of Subsistence And Settlement Citing McKinlay (1862), who was the first European to travel through the entire middle Cooper lake system, the region was densely populated at the time of contact. McKinlay's expedition, which was one of those initiated to search for Burke and Wills, involved the party spending several months in the lake system in 1861-1 862. During this time they were systematically examining each of the lakes and channel systems of the North- West Branch in their search for the missing explorers. McKinlay saw more than 700 people in the region during his travels, including more than 300 people camped along Hamilton Creek, immediately west of Lake Toontoowaranie in the Coongie system, 200-300 people along the North- West Branch just south of Coongie Lake itself, and at least 150 people around Lake Lady Blanche (1862: 37, 38, 46). Because of the length of time he spent in the region he was also able to observe how people utilised fluctuations in water levels across the lakes. McKinlay 's information shows that population densities were highest around the large, permanent waterholes on the channels of the Norlh-West Branch and the Cooper itself, and those lakes within the North-West Branch system which were currently holding water. Group composition appears to have been relatively fluid. Large aggregations of 1 50 or more were formed to take advantage of the flush of food resources which appeared as lakes began to fill. As they dried out again, people moved to other lakes, following the water, or split into small groups to harvest other resources or move on to other areas. Permanent or semi-permanent camps of between twenty and forty people were found along the permanent waterholes of the channel systems and this observation is also consistent with that of King, the surviving member of the Burke and Wills party who was cared for by Aboriginal people in the Innamincka region (Burke and Wills 1861). At these more settled camps habitations comprised substantial huts. Important food resources were fish, crayfish 78 E. WILLIAMS and mussels; waterbirds, and the variety of plant foods which germinated after flooding or local rain, especially nardoo. People favoured the lakes and rivers but after rain in the local catchment dispersed into small groups and travelled out into the drier dune country (McKinlay 1862: 49; Jones 1979). In these latter areas a flush of resources including small mammals, reptiles and certain types of plant foods, especially seed plants, appeared after the rain. Although there are no ethnographic observations of how the Cordillo system was used, it can be hypothesised that after local rain areas such as this became a focus for occupation, the water sources it now held resulting in a relative abundance of food. It is also possible that if semi- permanent pools of water remained for a considerable time throughout the year in the shallow channels between the lakes, some small groups of people might have occupied the region on a more permanent basis. While historical accounts mention the existence of the salt lakes there are no direct ethnographic observations of people using these areas. This implies that they were not favoured areas for occupation on a day-to-day basis. Dealing with periods of resource stress Although the Cooper lakes and channel systems contain a large range of food resources for much of the year, it appears that periods of resource stress were common, especially during droughts. Although there are no direct observations of the study area itself during these periods, historical material in Reuther (1981) provides information on strategies used by Aboriginal communities in the north-east of South Australia generally, to deal with the major fluctuations in environmental conditions characteristic of the region. Reuther noted that when resources were abundant, for example in periods after rain, communities utilised natural surpluses in a variety of ways. Large gatherings were organised to take advantage of the abundance of food and these aggregations had the primary function of satisfying social obligations through, for example, trade and exchange (1981; vol. 1: 42, 47^8). As well as these activities, people harvested more than they needed for immediate consumption, processing and storing the surplus for later use. Seeds, fish and grubs were the resources most favoured for this type of processing (1981 : vol. 1: 13, 47^18, 232, 273, 491; vol. 2: 547, 772, 835, 915, 927; vol. 4: 1852; vol. 7: 458, 549, 659). As well as these measures to maximise the productivity of a region, according to Reuther the distribution of food resources was also artificially managed. He recorded that rats and caterpillars were moved from one area to another so as to increase stocks in particular regions (1981 : vol. 2: 811-812; vol. 9: 232). Plant foods were propagated through the broadcasting of seed and the distribution of the branches of a particular vine species (1981: vol. 1: 107; vol. 7: 450). Areas of stream channels were dammed to increase stocks of fish (1981: vol. 8: 23). As well, certain localities were not utilised and were left in reserve for use during lean times, access to especially productive areas was strictly controlled and the boundaries of group territories were rigorously enforced (1981: vol. 1: 101, 123-128, 412; vol. 2: 559, 751, 1044; vol. 3: 1255, 1259). Transgressions, such as cases of people trying to obtain more food resources than their due, or attempting to accumulate resources, were severely punished according to Reuther, even by death (1981: vol. 1: 12-13, 123-129; vol. 2: 793-794). Further, instances were recorded of particular groups attempting to appropriate others' territory, sometimes successfully (1981 vol.1: 415; vol. 2: 896). During times of drought people fell back to those areas which had water, and fish and stored food resources were the main foods consumed (1981 vol. 1: 520; vol. 2: 547, 616, 758 ? 915, 950; vol. 2: 1259; vol. 4: 1740). Although Reuther doesn't specify these areas, it is likely that the Coongie system and the Cooper channel near Innamincka were used in this manner. Despite measures to counteract the effects of drought or a lack of surface water, local groups still suffered considerable stress. Reuther noted that many people perished at these times, from thirst, heat exhaustion or starvation (1981: vol. 1: 424, 535; vol. 2: 994-995, 1036; vol. 4: 1741- 1742, 1926, 2108; vol. 8:75, 118). Factors Other Than Food Resources Affecting Settlement It is a truism to state that it is not only the acquisition of food resources which affects the choice and nature of settlements. Access to raw materials necessary for the manufacture of goods and for trade, such as stone, wood, ochre, resin, bone, shell and the narcotic pituri, was another important determinant. Historical accounts show that the Innamincka region was an important centre both for the ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 79 production of technological items, especially stone artefacts and for a related reason, as a crucial node in the vast exchange network which spanned the continent (Kerwin and Breen 1981: 286 and McBryde 1987). Innamincka was especially renowned for its extensive grindstone quarries (see Reuther 1981 vol.2: 899-900, McBryde 1987). The quarries, other important sites involved in the extraction of resources and the manufacture of goods, and the exchange networks of the Lake Eyre Basin generally, are currently being studied by McBryde and are thus not reported on here. Although factors such as the utilisation of items for artefact manufacture and the location of trade and exchange networks are, as outlined above, crucial factors involved in settlement, not all of these features are equally archaeologically visible. For this reason I have chosen to look at one aspect that is more archaeologically visible than others, namely the characteristics of stone artefacts on lake-side sites and what these data, and other characteristics of the sites themselves, can tell us about Aboriginal use of the region. As outlined, stone suitable for the manufacture of artefacts is not uniformly distributed across the region and with the exception of Innamincka, stone sources and water sources do not generally coincide. Two main sources of stone arc present in the study area. Silcrete, chert and chalcedony are found as smallish pebbles on the surface of gibber plains, while better quality and more extensive areas of hard rock, comprising outcrops of chert, chalcedony, silcrete, quartzite and sandstone are found within the geological features known as the Innamincka and Cordillo domes (South Australia Department of Mines and Energy staff: pers. comm. and personal observation). Discrete areas of outcropping hard rock also occur within the linear dunefield south of the Cooper channel and east of Strzelecki creek (see Smith, Williams and Wasson 1991: 185). Although there are no historical accounts of how Aboriginal people in the region quarried and used stone, apart from Reuther's observations on the Innamincka grindstone quarries, field observations and geological data can be used to determine what may have been favoured sources. It appears that while gibber pebbles were utilised, this took place on a relatively limited basis compared to other sources because of the relatively small size of the nodules of stone that could be worked. Outcropping areas of hard rock were thus favoured over gibber pebbles because they provided larger fragments of stone and were more uniform in quality. As well, certain types of stone such as sandstone were not available as pebbles and were only located within hard rock localities. Using data from geological mapping of the region and advice from South Australian Department of Mines and Energy staff, it is estimated that the Coongie system lies at least 60 km from good quality stone sources while the Cordillo lakes lie at least 20 km distant. Particular stone sources which would have been favoured include: the silcrete (including chalcedony, cherts, cherty silcretes and quartzites) of the Cordillo surface which caps the mesas and cuestas of the Cordillo and Innamincka regions and also includes the silicified Eyre formation and the silicified Winton formation; the Eyre formation which includes sandstones suitable for grindstone manufacture, and also small pebbles of agate and fossil wood, and the Cadelga limestone which includes a cherty limestone. Surrounding these outcrops of hard rock and extending outwards for five kilometres or so are areas of gibber containing pebbles and larger blocks of stone eroded from these formations. Some of this material would have been utilised as well. A Model For Archaeological Site Distribution WiTHrN The Lake Systems Historical observations can be combined with information on the availability of stone to present a model of how the lake systems were utilised. The model makes a number of assumptions, namely that there is a relationship between the availability of water and food resources; and intensity of occupation as reflected by the size and nature of archaeological sites. Cutting across these factors are others relating to the availability of stone, the most common material found on sites. Using models developed by other researchers such as Hiscock (1987), it is hypothesised that there will be a general correspondence between the morphology of artefacts on a site and distance from the stone source. These models predict that as one moves significant distances from stone sources, characteristics of extreme reduction of stone material will be found in assemblages such as recycling, rejuvenation and rationing. Combining these different models, we can develop a unified set of hypotheses regarding the patterning of archaeological material. 80 E, WILLIAMS The Coongie system The Coongie lakes, with their abundance of food sources, and availability of water for at least part of the year, were areas with a high population density relative to other lakes in the region. Large and diverse occupation sites will be found in those areas which contain the most predictable and extensive areas of water. Such places include the permanent waterholes on the channels, the inlets to lakes, and those parts of the lakes that are deeper relative to the rest of the lake basin and thus hold water for a longer period of time. These latter areas are found on the southern margin, since as outlined above, the lakes in the region are slowly migrating southwards, cutting into longitudinal dunes on their southern shore. Their southern end is thus their deepest section. It is anticipated that given the known high population densities of at least the recent past, there will be a correspondingly high density of archaeological material (especially stone artefacts) on the large, favoured occupation sites. It is likely that on these sites people would not have moved as much compared to areas which contained less predictable resources. We can therefore expect evidence for the manufacturing and maintenance of tools such as stone artefacts. Given that the sites were used on a longer-term basis we can also expect a higher proportion of formal tool types than on sites with more unreliable water sources, and a greater proportion of finer- grained, good quality stone. As well, it is likely that a relatively larger number of artefacts will have edge damage from activities such as trampling, because of high population densities at the sites and residential stability (Scott Mitchell pers comm). Because it is possible that people from other areas may have been using these localities as refuges in times of resource stress, it is also expected that these places will contain a wider range of raw material than at other sites, including stone brought in from other regions. Because stone sources lay some considerable distance away (60+ km), primary reduction was unlikely to have been undertaken here and the stone may have been 'rationed'. Cortex, an indication of primary reduction, will be rare on pieces. Flakes and cores will be small in size and noticeably reduced. High population densities and length of site occupation may have exacerbated the incidence of rationing, so that these features will be particularly marked at the large sites near permanent or semi-permanent water. The salt lakes Given the absence of fresh water in these localities it is anticipated that they were not favoured as occupation sites. In times of high local rainfall it is possible that the lakes may have filled to some extent with rainwater run-off from the surrounding dunes. At these times salt levels may have decreased and they might have been used on a temporary basis as people pushed out into the dunefields after rain. The salt lakes may have had another role, as navigation aids for people travelling through the dunefield. They are often the only notable features present within extensive areas of dunes and are visible for some distance because of the sun sparkling on the salt surface. I conclude that the sites were used, albeit on a much more limited basis than the Coongie and Cordillo system. It is likely that groups using them were small in size and highly mobile. There is likely to be a correspondingly low density of artefacts, relative absence of formal tools, a more limited range of raw material types present and a lower proportion of good quality fine-grained raw material on occupation sites. Little manufacture or maintenance of artefacts will have been carried out on site. Given that the lakes are often well within the dunefield and thus some distance from stone sources, the stone discarded here will have been noticeably reduced. The Cordillo lakes With their better resources than the salt lakes the Cordillo lakes are likely to have been the subject of more intensive occupation but on a less pronounced basis than the better-watered Coongie system. It is anticipated that sites will be smaller in size than in the Coongie and fewer artefacts will be present. Because people occupied these sites for shorter periods of time it is expected that there will be less evidence for the manufacturing and maintenance of stone artefacts. A lower number and a smaller range of formal tool types will be present, and there will be less variation in raw material type. Larger sites are more likely to be located in areas of more predictable water and food resources such as the inlets to lakes or the small waterholes on channels. Given that the Cordillo lakes are closer to raw material sources than the Coongie it is expected that stone artefacts will not be reduced as much. Cores will be more common and cores and flakes ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 81 will be heavier and larger in size, on average. A higher proportion of flaked pieces will contain cortex. The Archaeological Surveys: General Archaeological Features Of The Lake Systems Before presenting material on the archaeological sites examined as part of the study I will outline some of the factors affecting the location and recording of archaeological material in the field. The visibility of archaeological material, especially stone artefacts, is reasonable compared to many other areas. The aridity of the region means that dense vegetation is often absent and much of the ground surface is exposed. This also means that stone sources are relatively straightforward to detect. Open sites are the major archaeological site type in the region because rock formations which form rock shelter sites are extremely rare here. Only one shelter (a very small and shallow one located near Innamincka) is known for the whole region. Although open sites arc the major site type available for analysis, given the good ground surface visibility this is not necessarily the problem that it is in other regions. Stratified sites are extremely rare nevertheless and it is often very difficult to determine whether archaeological material is unequivocally in-situ, has been deflated down from other units, or otherwise re-deposited. I will present general features of the archaeology of the lake systems first and compare them with my model. In the following sections I will discuss more detailed analyses of assemblages from a sample of collected sites so as to make a comparison between localities within the region. The Coongie system The results of this work are presented in Williams (1988) and are briefly summarised here, along with other material from 1987 surveys. Before outlining this information it is noted that not all of the lakes in the Coongie system were surveyed for archaeological sites. Coongie itself was not examined because it was already being systematically surveyed by a team of archaeologists from the South Australian Aboriginal Heritage Branch, (now within the Department of State Aboriginal Affairs), in relation to the development of a management plan for the region. The results of that study are not yet available for analysis. Lake Goyder was not examined because in the years which formed the subject of my fieldwork (1986-1987) the lake was flooding out and thus conditions were not conducive to archaeological survey. The most common site type around the lakes surveyed for the study comprised scatters of stone artefacts. In all sites surveyed these scatters were either located within upper recent sediments or appeared to have deflated down from such units to sit as a lag on more compacted surfaces. On many of the sites, scatters of freshwater mussel shell were also present. Although not all the margins of every lake were examined, it appeared that stone and shell was present as an intermettent scatter intermittently around the margins of lakes that were surveyed (Toontoowaranie, Marroocutchanie). These were often present as small scatters of shell with numbers of artefacts, similar to Meehan's (1982) 'dinnertime camps'. Artefact density and site size increased as one approached permanent water sources and the largest sites were found in the areas outlined earlier - permanent waterholes on the channels, the inlets to lakes, and the deeper, southern margins of the lakes. These latter sites often incorporated extensive scatters of shell as well. Compared to the large sites, site size and density of material was comparatively low on the pale- coloured transverse dunes. On these sites stone artefacts and scattered termite mound hearth fragments were the most common form of material. Typologically, most, if not all artefacts dated to the mid to late Holocene. The main types noted were tula adzes and adze slugs, points, backed blades, small scrapers, cores, flakes and fragments of large, flat grindstones of the type described by Smith (1986). As well, single examples of large, cube-shaped silcrete cobbles which were commonly ground on one or more surfaces and in some cases also had flakes removed from edges and had thus been used as large cores, were often also present. Raw materials favoured for artefact manufacture were chalcedony, chert, cherty silcrete, a coarser-grained silcrete, and quartzite. Other artefacts such as edge-ground axes and flaked pieces of quartz crystal were also found but appeared to be restricted in distribution to the lakes most difficult to access by vehicle. It appears that such artefacts were formerly present on other lake-side sites but, together with other types of artefacts such as complete grindstones, have been removed by collectors (a conclusion supported by oral tradition in the region and also personal 82 E. WILLIAMS observation of collections held by individuals resident in the region). Apart from the occasional large silcrete cobbles and remaining grindstones left on sites, stone artefacts were mostly small in size and appeared to be noticeably reduced. As well, although density of stone material varied with location, even on sites with comparatively higher numbers of artefacts, density was still relatively low compared to artefact scatters found in many other parts of the continent. On the larger sites average density for collected sites was of the order of 6-8 artefacts/ square metre while on the smaller sites maximum densities were only 2-A artefacts/square metre. Occasionally larger flakes and horsehoof cores were found on sites or located as isolated finds within the dunefield or near lakes. There appeared to be no consistency in their location though and these were often in association with artefacts that typologically dated to the mid to late Holocene. As well, they were not noticeably patinated or weathered, and did not appear to be derived from Pleistocene sedimentary contexts, suggesting that most of these are mid to late Holocene in age. Hearth material, usually fragments of burnt termite mound, was often scattered across sites and isolated hearths were also present. The amount of hearth material rose in density on the larger sites and on these latter places fragments of bone (mostly that of fish and small mammals) were often found as well. The large sites are indeed extensive - examples up to 10 000 square metres and 7 000 square metres in extent respectively are described in Williams (1 988: 59). Dating of material on sites confirms that implied by the typology of artefacts, that is, of the late Holocene. Details of radiocarbon dates are provided in Williams (1988). The archaeological data is consistent with the ethnographic observations that the Coongie lake system was extensively used and densely populated. The presence of good quality raw material and a range of artefact types, including formal tools and manufacturing debris, suggests that residential stability could have been fairly high in comparison to other areas. This is also supported by the wide range of archaeological material present, including bone and shell. The small size and marked reduction of artefacts indicates rationing of raw material was taking place and this is consistent with the distance of the area from stone sources. The salt lakes As expected, within the areas sampled for sites for Coorie Coorie Tillie and Lake Deception, archaeological material existed in very low densities. At Coorie Coorie Tillie for example, along the 5 km stretch of the transverse dune on the northern margin of the lake only five isolated artefacts were seen: three amorphous flakes, one tula slug and one small core. All artefacts were manufactured from silcrete. Archaeological material was similarly rare at Lake Deception. Sites comprising occasional small, discrete groups of artefacts were found in the vicinity of both lakes but not within the transverse dune systems. Instead they were set back one or two longitudinal dunes from the lake margin, on the surfaces of small claypans. The sites found in these areas comprised small (mostly less than 30 artefacts), discrete scatters of stone artefacts. Density was very low, much less than 1 artefact/square metre. Neither mussel shell nor animal bone was present on these sites or on other parts of lake margins but occasionally examples of isolated termite mound hearths were found in association with the stone artefact scatters. One of the termite mound hearths at Coorie Coorie Tillie was dated. To examine particular features of the region's environment which may be affecting dates, the material was separated into a humic and non-humic fraction and each sample separately dated. The humic fraction returned a date of 1150 180 BP (ANU 6181) and the non- humic: 710 170 BP (also ANU 6181). The disparity in the dates can be explained as follows. John Head, ANU Radiocarbon Laboratory, confirms that the non-humic fraction, comprising clay particles, binds to younger carbon and clay material, biasing the date and making it too young. To account for this the non-humic fraction, which although comprising polymers which can incorporate younger carbon, was treated with solvents so as to extract the older carbon which is more likely to be reflecting the true age of the sample (pers. comm.). The older carbon is dated and the result, 1 150 180 BP, is the best estimate of the sample's true age. Apart from the differences in density and site composition from the Coongie sites, other distinguishing features were also apparent. Most of the artefacts were manufactured from a relatively restricted range of raw material, assemblages being dominated by a grey, comparatively coarse-grained silcrete. Some cores were present and also occasional formal tool types such as tulas, scrapers or points. Artefacts appeared to be noticeably larger than those at ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 83 Coongie. Because of the low density of material, and also given the difficulty of access to the sites, collections of artefacts were not made when the sites were visited and so a more formal comparison between these features and other lake systems is not possible. The low density of material present on these lakes suggests that while they were utilised, they were not favoured for occupation. The restricted range of raw material found, its contrast in quality to the fine-grained material found on the Coongie system and the relative absence of manufacturing debris, indicates people used these areas on an opportunistic basis. The low numbers of artefacts present also suggests that group sizes were relatively low. The Cordillo lakes Vehicular access to the lakes was difficult because of their remote location and thus only one lake. Koonoomoorinna, was surveyed in detail. Reconnaissance trips to other lakes in the region revealed that the general patterning of sites on this lake is similar to that on others. As for other parts of the study region, the most common site types seen were scatters of stone artefacts. Formal tool types were similar to those observed on the Coongie sites, comprising tulas and other adzes, scrapers, points, fragments of grindstones and the large silcrete cobbles with ground and/or flaked surfaces. Formal types did not appear to be as common within assemblages as for the Coongie and no shell or bone was present on sites. The range of raw material did not seem as extensive and Finer grained raw material such as chert and chalcedony was found in lower densities. Artefacts appeared to be larger in size and not as heavily reduced. Hearths were also not as common and no hearth sites were observed on any of the sites ringing the immediate margin of the lake. A small number of hearths were found in the inter-dune corridors one or more dunes back from the lake margin. One, comprising burnt calcrete nodules, was dated and returned a 'modern 1 radiocarbon age (100.2, 1.8 %M, ANU 6183), although this sample may have been contaminated by younger carbon from the surrounding locality (John Head pers. comm.). The densest, most extensive scatters of artefacts were found in those areas with the most reliable sources of water - inlets, the southern margin of lakes, and on the edges of possibly semi- permanent waterholes along the channels which connect the lakes. A large scatter along one of the channel waterholes contained a number of hearths. Intermittent, low density scatters of artefacts were found along the lake margins and on the transverse dunes to the north. The archaeological patterning is consistent with the hypothesis that the Cordillo lakes were utilised more than the salt lakes but less than the better- watered Coongie area. This is supported by the fact that the Cordillo sites contained a smaller range of archaeological material than the Coongie sites (no shell or bone, hearths not as common, a lower density of formal tool types and of better quality stone). The relative lack of reduction of artefacts could indicate either less residential stability or a closer distance than Coongie to raw material sources, or both. Now that I have presented general observations of the archaeology of the lake systems I will outline the results of the more detailed comparisons between sites. This work provides a more fine-grained analysis of differences that may exist across lake systems and between systems. More Detailed Comparison Of Assemblages From Individual Sites Because of the relatively low density of artefactual material within the region, for reasons of statistical analysis I only made formal collections from large sites. The sites analysed in this way are outlined below, along with the results of the analysis. Before discussing the sites it is noted that although some of the analysis of the material was published in a preliminary form by Smith, Williams and Wasson (1991, see for example Tables 1 1 and 12), a number of the figures in that paper vary slightly from those presented in this publication. This is due to the fact that the results of the analysis of the Coongie sites outlined in the 1991 paper were only at a preliminary stage. The material presented in the current paper reflects the completed analysis of the assemblages. Sites chosen for analysis Five sites were chosen for analysis: Marroocoolcannie 2 (MCO 2), Toontoowaranie 1 (TTW 1), Lake Lady Blanche midden (LLB), Koonoomoorinna inlet (KOI) and Koonoomoorinna second white transverse dune (KSW). The first three sites are located on Coongie lakes - their name indicates on which lake they are situated, and the latter two sites are located on the southern-most of the Cordillo lakes, Koonoomoorinna. All sites comprise large, discrete scatters of stone artefacts, together with E. WILLIAMS other material such as shell and bone in some cases. The lakes were chosen for sampling and analysis to determine whether there were any major changes in stone artefact manufacture and discard across the Coongie system, Each site was chosen because it was the most extensive seen for the sampled areas of that particular lake. MCO 2 lies along the western margin of the channel feeding from Lake Marroolcoocannie to Lake Marroocutchanie. It covers an area of approximately 2 600 square metres and comprises a scatter of artefacts and shell deflating down from the uppermost recent unit of a pale, longitudinal dune. The material exists as a lag on a blow-out within the dune. At the western end of the site is a localised area of burnt shell and bone, the latter comprising the remains offish and small mammal species. A hearth, which lies on the eastern part of the site, dated to 1 130 1 10 BP (ANU 429). A baseline was laid through the middle of the scatter so as to bisect both the area of burnt bone and the hearth. All artefacts along a strip 60 m along the baseline and 1 m either site of it were collected for analysis. TTW 1 lies further to the north than MCO 2, located on the southern margin of Lake Toontoowaranie. It covers an area of 7000 square metres and comprises stone artefacts, shell and the remains of the collapsed, wooden framework of a domed hut, all lying on recent sediments on the top of a longitudinal dune. A sample of shell from the site returned a date of 330 80 BP (ANU 5425). A 20 x 2 metre grid was laid out within the centre of the scatter and all artefacts within this were collected. LLB lies to the north and west of MCO 2 and TTW 1, on the inlet channel to Lake Lady Blanche, one of the terminal lakes in the Coongie system along with Lake Sir Richard. These two lakes hold water for a shorter period of time than those closer to the North-West Branch. When the lake was visited in August 1987 it was dry and had a salt crust on the lake floor. This was in contrast to the other Coongie lakes which were all full of water. The site comprises a large scatter of stone artefacts, shell and bone extending for 12 000 square metres, eroding out of the uppermost units of a recent, white transverse dune associated with the inlet channel and flood-out zone of the lake. The bone comprises fish bone and golden perch otoliths. It is the largest and richest site in the inlet area. A 15 x 15 metre grid was laid out on the eastern edge of the exposure, where the densest scatter of shells, otoliths and artefacts was eroding out of the sand. All artefacts within the grid were collected. The relative lack of water in the lake compared to others in the Coongie system reflects the distribution of sites in the area. Large sites such as this one, which are not only extensive but contain a variety of material, are restricted to the margins of the inlet channel. Isolated scatters of stone artefacts are found on the surface of the complex of at least five separate transverse dunes lying to the north-east of the lake, but these sites are small, discrete scatters of stone artefacts and do not contain shell, bone or hearth material. Although the lake contains water on fewer occasions than others in the system, it was much favoured by Aboriginal occupants. McKinlay visited the lake in January 1862 and found it full of water, * between five and six feet deep, and seven and three quarter miles in circumference ... and tens of thousands of pelicans on it ... with innumerable other birds ... and plenty of fish* (1862: 38). There was a large camp of at least 150 people on the lake when McKinlay was there and he observed that fish and 'nardoo 1 was their chief sustenance (1862: 42). For the Cordillo lake, Koonoomoorinna, two areas were sampled for artefacts. Both sites, KOI and KSW were the most extensive discrete scatters of artefacts seen within the vicinity of the lake and this is why they were chosen for analysis. KOI is a very extensive scatter of artefacts which follows the inlet to Lake Koonoomoorinna for several hundred metres. The site comprised stone artefacts only - no shell, hearths or bone were present. A small part of the site (10 m by 2 m) where the densest collection of artefacts was found, was gridded and a collection made. KSW comprises a discrete scatter of artefacts lying as a lag within a blow-out in the second transverse dune north of the lake. Like KOI, only stone artefacts were present within the site. MCO 2 and TTW 1 lie 65 km from sources of outcropping hard rock in the Innamincka dome and 60 km from the Cordillo dome sources. LLB lies 40 km from the Cordillo dome sources and 65 km from the Innamincka dome. KOI and KSW lie 30 km from the Cordillo dome and 90 km from the Innamincka dome. For comparative purposes, the five lake-shore sites are compared with each other and also with one from the waterless core of the dunefield, site JSN. This latter site, which is reported on in Smith, Williams and Wasson (1991), comprises a scatter of stone artefacts and hearth material, and ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 85 Table 1. Characteristics of artefacts within assemblages, excluding ground pieces Site mean weight % flakes % cores % formal in gm, (sd) tool types Table 3. Mean weights of cores (gm) Site mean sd KSW 7.55 (17.57) 70.4 3.4 3.8 n=291 KOI 5.14(20.53) 48.4 4.2 2.6 n=308 LLB 2.65 (7.11) 45.2 0.6 11.9 n=336 TTW1 1.74(3.33) 59.5 — 4.2 n=264 MC0 2 2.17(8.32) 50.1 0.6 5.6 n=661 JSN 6.6 (sd not 37.3 2.7 not n=407 available) available is the only known Pleistocene site in the region. Hearths from the locality date to around 14 000, 10 000 and 2 500 BP. Associated with the hearths and lying as a lag on two inter-dune pans within longitudinal dunes is an extensive scatter of artefacts. Despite the Pleistocene dates for the hearths, analysis of the stone artefacts strongly suggests that they are late Holocene in age (Smith, Williams and Wasson 1991: 184). JSN was chosen as a comparative site for the lakeside sites because it comprises the only known extensive scatter of stone artefacts for the core dunefield region. Most core dunefield sites only consist of an isolated artefact or very small scatters of stone artefacts (personal observation). The large size of the artefact sample (N = 407) from JSN and the fact that the artefacts appear to be late Holocene in age, similar in age to the lakeside sites, makes it the only comparative material currently available for the Coongie and Cordillo sites. KSW KOI LLB TTW1 MC0 2 JSN 10 58.29 50.49 13 42.62 81.75 2 13.7 1.90 4 7.65 6.56 12 89.2 57.9 % artefacts with cortex 12.4 11.0 4.29 9.2 3.2 20.0 Table 4. Percentage of artefacts larger than 1 cm in length which contain cortex. Site KSW KOI LLB TTW 1 MC0 2 JSN Analysis of the sites Broad characteristics of the stone artefact assemblages are presented in Table 1. As predicted by the model, the Coongie assemblages have features reflecting an extreme degree of reduction. Artefacts (all pieces of stone on the site worked in some way, excluding ground pieces) are extremely small in size (measured by mean weight in grams) and cores are very rare, or even non-existent in the case of TTW 1 . The Coongie sites are very similar to each other but are different to both the Cordillo sites and JSN, which contain larger artefacts and a greater number of cores respectively. The differences in mean weight between the Table 2. Values of t weights of artefacts for comparison between mean Table 5. Raw material composition of the assemblages (excluding small shatter < 1cm in length and including ground pieces) Site KSW KOI LLB TTW 1 MCO 2 Site % Finer % Coarser % Very coarse KSW KOI not - - grained grained grained (eg sandstone) significant - — — KSW (n-278) 47.1 41.7 11.2 LLB t = 4.71 t = 2.11 - - - KOI (n-239) 37.2 59.0 3.8 TTW1 t = 5.33 t = 2.72 not LLB (n=223) 72.6 19.3 8.1 significant - - TTW 1 (n-223) 66.4 27.4 6.3 MC0 2 t = 6.48 t = 3.22 not not MC0 2(n=659) 59.8 23.7 16.5 significant significant JSN(n-401) 59.5 33.2 7.4 86 E.WILLIAMS Table 6. Mean weights and mean lengths of complete flakes for different raw material types. Standard deviations are shown in brackets. Note: Only those artefacts > 1 cm in length have been included in this analysis because of the difficulty in determining whether smaller pieces were derived from finer or coarser-grained raw material. Ground pieces and sandstone have also been excluded from the sample. Site mean weight (gm) of complete flakes mean weight (gm) of debitage>l cm in length Mean length (mm) of complete flakes finer coarser finer coarser finer coarser grained grained grained grained grained grained KSW 3.9(4.1) 12.0(18.6) 3,2 (2.7) 7.4(14.1) 25.0 (9.4) 31.7(13.6) KOI 2.3 (2.6) 6.71 (15.6) 2.3 (2.9) 5.0(7.1) 17.8(7.1) 25.2 (14.60) LLB 1.3(2.0) 2.1 (2.8) 3.1 (4.6) 9.3 (21.8) 15.8(5.1) 19.1 (6.8) TTW1 0.9(1.1) 4.8 (6.4) 1.0(1.7) 7.4(14.9) 15.8(6.1) 26.3 (14.7) MC0 2 0.9(1.5) 8.0 (24.3) 0.9 (0.9) 3.5 (5.3) 15.2(5.3) 23.5(16.2) JSN 3.1 (3.2) 4.7 (10.6) 5.6(13.0) 3.5 (6.4) not available Coongie and Cordillo assemblages are significant at the 0.05 level (Table 2). The figures for JSN are not available for comparison but it is anticipated that if they had been available, the mean weight of the assemblage would be statistically different to the Coongie sites but of the same order as the Cordillo assemblages. As predicted by the model, the three groups of sites are very different in relation to the reduction of cores (Table 3). The cores from the Coongie sites are extremely small and, as outlined in Smith, Williams and Wasson (1991:187), very close to the theoretical limit at which bipolar techniques become necessary for further reduction. Consistent with this is the observation that in a further attempt to ration raw material, worked-out or small, broken cores have been recycled into small scrapers in the Coongie assemblages (N=9 S mean weight = 8.2 gm, sd = 4.9). The Cordillo sites contain cores that are noticeably larger and heavier than those on the Coongie sites, while JSN contains on average larger cores again. There is no evidence that cores have been recycled as implements at either the Cordillo sites or at JSN. The Cordillo sites contain proportionately greater number of flaked pieces with cortex than Coongie (Table 4), which is as predicted given that they lie closer to raw material sources, although the figures for TTW seem high in comparison to the other Coongie sites. The JSN assemblage contains an unusually high number of flakes with some degree of cortex. This feature is unexpected and as well as the relatively larger size of cores and flakes, indicates that the JSN assemblage shows neither the effect of relatively large distance from stone sources, nor mediating responses offsetting a diminishing stock of stone (see Smith, Williams and Wasson 1991: 187). This issue is discussed in more detail later in this paper. Regarding raw material composition for assemblages, Table 5 shows that the Coongie sites contain proportionately more finer grained material (classified here as chalcedony, chert and cherty silcrete) than either the Cordillo sites or JSN and this is as expected (coarser-grained material has been classified here as granular silcrete and quartzite). Table 6 shows that flakes and debitage from fine-grained material arc on average smaller in size and lighter in weight than those from coarser-grained material, but this docs not explain all the variation in size between assemblages between sites in different areas. Fine- grained flakes and debitage from the Coongie sites are in virtually every case smaller in size and lighter in weight than fine-grained pieces from either the Cordillo sites or JSN. As well as these trends, the Coongie sites also show a wider range of raw materials as predicted. MCO 2 has 13 major types of raw material, TTW 1 has eight types, LLB has seven, KOI has five and KSW has seven. In terms of the artefact analysis it was decided not to differentiate artefacts with retouch from those where this feature appeared to be absent. This is because of the difficulty found in unambiguously identifying examples of retouched pieces within the Coongie and Cordillo assemblages. Numbers of artefacts within the assemblages contained what I recorded in my analysis as 'edge damage', where I describe this as an edge where very small flakes had been ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 87 removed and there is also a gloss or polish (visible under low-power magnification). However given the small size and thin cross-section of many of the flaked pieces it was impossible to determine whether this edge damage had occurred as a result of use, or as deliberate retouch, or from some other process such as flaking or trampling for example. Accordingly it was not possible to examine one of my hypotheses, that sites near more permanent water sources would show more edge damage from trampling. Because of the difficulties in identifying what had caused the edge damage, none of the statistics for this feature have therefore been listed for this analysis. It is noted that in the assemblages examined there were very few artefacts with pronounced retouch (where this involved larger flakes being removed along an edge and/or very obvious polish) which did not also comprise a formal tool type such as a scraper, adze, point or backed blade. For the very few artefacts which did not fit in to a formal category they were nonetheless listed under 'formal types' in my tables. Given that this classification methodology differed from that used for the JSN artefacts, it has not been possible to compare the two groups in a formal sense. Nonetheless some comparison are possible. In terms of my category of 'formal tool types', the Coongie sites contain a greater proportion of these artefacts than the Cordillo sites (Table I), as predicted. Adzes such as tulas and tula slugs formed the largest proportion of formal types in the Coongie sites (59% for MCO 2, 45% for TTW 1 and 55% for LLB). The remaining implement types comprised scrapers, points and backed flakes. For the Cordillo sites in contrast there is greater variation in the formal types of stone tools represented on sites. Adzes comprised 36% of the formal types present at KOI but only 18% of those at KSW. At KOI scrapers comprise the most common formal type (50%) while at KSW, points, backed flakes and backed blades are the most common type at 64%. Regarding the artefacts comprising formal types, it was expected that examples will be more heavily reduced at sites near permanent water. Examining the largest sample of a formal type available, tula adzes and tula adze slugs, it is found that this is not the case. There is no clear trend in mean weights between the Coongie sites (MCO 2: mean - 8.5 gm, sd - 5.02, n = 15; TTW 1: mean = 3.25 gm, sd - 1.35, n = 2; LLB: mean - 14.18 gm, sd = 1 0. 1 , n = 1 9) or between the Coongie and Cordillo sites (KOI: mean = 3.7, sd= 1.0, n = 2; KSW: mean - 6.8, sd = 2.7, n = 2). This suggests a number of factors. While overall characteristics of the assemblages presented in Table 1 show that stone was being curated and rationed in a very intensive way for the Coongie sites, the variability in size at discard for tulas and tula slugs, especially the relatively higher weights for the LLB site, suggest that occasionally the site occupants were not behaving as if raw material was as scarce as it appeared to be. This suggests that while certain factors such as the balancing of procurement of food and water sources with that of raw material is responsible for much of the variation seen on sites, people did not always behave as if this was the only equation affecting the nature of occupation and discard of stone artefacts. Although the JSN assemblage has been classified according to a different typology, a preliminary analysis of it reveals that formal types are rarer than at either the Coongie or Cordillo sites. This is consistent with the model. As well, it is further observed that adzes are not present on the site. A possible explanation for this relates to the lack of hardwood trees within the core dunefield region supplying suitable wood for adzing activities. The absence of such trees is very likely related to the availability of water - hardwood specimens being found on the banks of waterholes and lakes. Comparison Of The Trends In TirE Assemblages With Data From Other Parts Of The Arid Zone The trends seen in the assemblages for the region are also echoed in other arid zone work available for comparison. The major study which provides comparative data is that of Veth (1993) for the Western Desert region, Western Australia. His analysis shows similar patterning to that outlined above, with features of assemblages from sites at permanent waters indicating more intensive stone reduction than for cither semi- permanent waters or ephemeral water sites. His data show that differences in intensity of reduction are not as clear between semi-permanent and ephemeral water sites as they are between these sites and more permanent ones (see 1993: Tables 8.21-8.31). For a number of features, ephemeral sites show trends characteristic of more intensive reduction than semi-permanent sites, which is not predicted by Veth's model. This appearance of 88 E. WILLIAMS conflicting trends also appears to be echoed by comparisons between the Cordillo sites and JSN. The sites are more similar to each other than they are to the Coongie sites but there is much variation in the nature and direction of trends between the two groups. This suggests that the permanent sites, in Veth's sample and mine, form clear groups but other sites, while showing some similar trends, are more variable. The association of higher proportions of formal types in assemblages, especially adzes, with more predictable water sources, is also a trend noted by Veth (1993: 95-96). Veth also found that there was some patterning in the amount of reduction of tula adzes and relation to water sources but that there was a considerable degree of variation (1993: 98-101) and this is also found in my samples. In his discussion of his results of the artefact analysis, Veth concludes that the timing and intensity of site use are the major factors responsible for variation in the assemblages, rather than distance to raw material sources and the operation of specific discard criteria (1993: 101). This also seems to be the case for the north- east South Australian sites, as I will outline below. Major Factors Affecting The Patterning Of Sites In The Region One of the predictions of my model of site patterning is that apart from water sources, availability of stone sources is be one of the most important factors responsible for trends seen and that Aboriginal occupants favoured raw material sources closest to the site. An examination of the JSN assemblage reveals that the people who visited the site did not behave in this manner. Rather than using the closest stone to the site (40-50 km to the east) for artefact manufacture and discard, the sourcing of stone artefacts from the site indicates that they instead used stone from Della-Dullingari which lies 115 km away, to the north-east (Smith, Williams and Wasson 1991: 185). Despite the use of such distant material, the assemblage does not have features characteristic of extreme reduction (Smith, Williams and Wasson 1991: 185-189). Cores have not been noticeably curated and have not been recycled as other implements. Mean weight of artefacts is significantly greater than for the Coongie sites, despite the fact that utilised stone sources are considerably further away than for the Coongie. Flaked pieces contain much higher proportions of cortex, also consistent with less intense reduction. The patterning of the JSN assemblage is characteristic of a situation where raw material is closer rather than further away. Yet the paradox can be resolved if we instead see stone source utilisation in terms of re-supply time rather than just in terms of linear distance of stone source from occupation site. For example, if people stopped at JSN as part of a journey to the Della- Dullingari region, they would have been aware that they were only a couple of days away from being able to replenish stone sources. As well, it is likely that the duration of visits to the JSN were relatively short, so that the demand for stone was easily met by the stock-in-hand without recourse to the recycling and the extreme reduction of stone evident at the Coongie sites (Smith, Williams and Wasson 1991: 188). The JSN site shows that it is not necessarily linear distance from the closest stone sources that is a major factor in the patterning of stone artefact manufacture and discard. The key factors are instead the length of time spent at the site, which in turn relates to the length of time before the group could replenish stone sources. This line of reasoning can be applied to the Coongie sites as well. The extreme reduction of stone reflected by the assemblages can be seen as the archaeological correlate of residential stability within the lake system relative to other environments in the region. It appears that people chose to stay closer to the lakes even if it meant that they were restricting their access to stone, thus triggering what appears to have been significant pressure on the way they used the resource. This is despite the fact that they could have chosen to be more mobile, travelling more often between stone sources and the lakes. In this latter case, features of such extreme reduction would not be visible in assemblages. Therefore, like Veth for the Western Desert region, I see the main determinant of site patterning in the Coongie and surrounding areas as relating to the timing and intensity of site use rather than to distance from raw material as such. This indication of relative residential stability in the Coongie, combined with the high population densities observed in the area at the time of contact implies that the system may have been vulnerable to significant stress given the extreme environmental fluctuations characteristic of the region. It is likely that the system worked ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 89 smoothly when resources were abundant but in cases when an anticipated flow of fresh water failed to come down the Cooper or when there was widespread drought, it is possible there were major problems. The fact that the current situation where the Coongie captures most, if not all of the Cooper flow, is potentially very unstable and channel flow could be completely switched on and off as a result of only slight environmental shifts, would exacerbate this. Ethnographic observations support the notion that stresses were operating on the system. The rigid sanctions outlined earlier involving the enforcement of group boundaries and access to resources indicate this, along with the information that many people in what is now the north-east South Australian region perished during droughts. Further, Wasson's observation of a dune mobilisation event in the late Holocenc that might be linked to the intensity of the occupation of the region, is another feature that suggests the system might have been under some stress. Evidence for 'intensification '? It is tempting to suggest that the characteristics of the mid-late Holocene occupation of the region reflect an intensive settlement pattern and are derived from a process of 'intensification' (see also Williams 1988: 61). For example, all of the dates for occupation sites for the Coongie system fall within the late Holocene and on all sites recorded within the region, including JSN, the typology of artefacts suggests a mid-late Holocene date. While it is tempting to conclude the region does provide evidence consistent with the operation of such a process, as in my 1988 paper I am cautious about maintaining that these features provide definite evidence. I take this position because it is likely that the environmental instability could be causing major biases in the archaeological record. Such instability could be selectively removing older sites from the archaeological record, or rendering them archaeologically invisible through sedimentation, or even resulting in vast fluctuations in channel flow thus affecting the filling of lakes. I therefore conclude that while the evidence for intensification is persuasive, further geomorphological work is needed to more fully expand on factors affecting the visibility of archaeological material. Given these issues, can any conclusions can be drawn about the occupation of the arid zone, especially that of the more distant past? Early Occupation Of The Arid Zone 1 will conclude the paper with some brief and speculative comments about models relating to when the arid zone was first occupied on a consistent basis. As noted in Smith, Williams and Wasson (1991: 190) the evidence from the JSN site indicates regular use of the dunefield region during the late Pleistocene, when there was probably also occupation of the riverine corridors. The presence of Velesunio shell at JSN in a Pleistocene context is tangible evidence of some link with these riverine habitats. The nature of the archaeological material present on the site suggests that late Pleistocene use did not involve much on-site stone working and that this changed some time during the mid-late Holocene when visits to the area appear to have become more prolonged and to have involved the grinding of seeds and the manufacture and maintenance of chipped stone artefacts (ibid). Dating of hearths to the period around 12 000 BP for hearths on the lower reaches of Coopers Creek (Veth, Hamm and Lampert 1990) is also consistent with late Pleistocene occupation of what is now north-east South Australia. Despite the presence of the dunefield site of JSN, Veth (1993) has developed a model suggesting that the continent's core arid zones were only colonised as recently as 5 000 years ago. Veth maintains that the evidence for Pleistocene occupation of JSN does not disprove his model since the Strzelecki dunefield region is located within an area of co-ordinated drainage, thus implying that it was not part of the core arid zone (1993: 111). Veth's model is an interesting one, but it could prove difficult to test. His 'core arid zones' comprise the deserts of the Great Sandy Desert, Great Victoria and Simpson Deserts. These regions are by definition sandy deserts and are thus areas where outcrops of hard rock are rare and therefore rock shelters with potential Pleistocene sediments are either non-existent or uncommon. Rockshelters such as those examined by Veth (1993) in the Gibson Desert, have late Holocene occupation resting directly on bedrock and thus do not provide contexts for the examination of occupation (or the lack of it) for the period before the late Holocene. The fact that the presence of rockshelter sites which preserve Pleistocene sediments is so crucial for the examination of whether Pleistocene occupation took place or not, is shown by Veth's 90 E. WILLIAMS discussion of his model (1993: Chapter 9). All of the eleven Pleistocene sites and early Holocene sites which lie a considerable distance from co- ordinated drainage, and thus form the basis of his model, are rockshelters. Knowing that rockshelter sites which preserve Pleistocene sediments are likely to be extremely rare in the regions important for Veto's model, it is also unlikely that Pleistocene open sites will be easy to locate. Extensive outcrops of hard rock are not common in the sandy deserts and thus stone suitable for the manufacture of artefacts is also a scarce commodity. It is therefore likely that stone artefact manufacture, one of the maj or archaeological indicators of human occupation, has to take place on a relatively intensive basis before it becomes archaeologically visible in these regions. The JSN site for example indicates that Pleistocene occupation is not as archaeologically visible as regards stone artefact discard as is mid-late Holocene occupation. This evidence is complemented by my findings for the study area that artefacts do not appear to be present within visible Pleistocene contexts. This is despite that fact that we know from the evidence of JSN that people were using the region during the late Pleistocene. Further, the logistical difficulties of carrying out systematic archaeological surveys for sites in remote locations such as sandy deserts, compound these problems. JSN was not found as the result of systematic archaeological sampling but was instead located virtually by accident during the course of a geomorphological study of the dunefield by Wasson. The question of when the core arid zones were first occupied on a systematic basis thus remains a challenging one. Until this question is resolved, JSN remains the continent's sole example of a Pleistocene site located within the midst of the continental dunefield, albeit a dunefield in the vicinity of co-ordinated drainage. The spur to the finding of Pleistocene sites within Veth's 'core arid zones' will be to expand further on the nature of the continent's Pleistocene occupation and to also determine the antecedents of the patterning seen in the mid-late Holocene. Acknowledgments The project reported on in this paper was funded jointly by the National Research Fellowship Scheme and the Department of Prehistory, Research School of Pacific Studies, (now the Division of Archaeology and Natural History), the Australian National University. I thank these bodies for support. The Australian Heritage Commission also provided some assistance and I thank Betty Meehan particularly in this regard. Bob Wasson was instrumental in encouraging me to initiate the Coongie project and I especially thank him for his advice and support during the course of the work. SANTOS provided much logistical support for fieldwork and I particularly thank Oleg Morozow and Steve Tunstall for their assistance. The Aboriginal Heritage Branch gave permission for collection of sites and the excavation of JSN. Chris Dodd (then Aboriginal Liaison Officer, SA NPWS) gave approval for the excavation of JSN and I especially thank him for his assistance with work on the site. I wish to thank all of the people who provided assistance in the field: Doreen Bowdery, Nick Drayson, Ros Fraser, Parry Kostiglou, Gary Dunnett, Pam and Colin McDonald, Ben Smith, Barney Stevens, Steve Sutton, and Keryn Walsh. I am also indebted to Barry Saunders and Vince for their advice and support in the field, and the Royal Flying Doctor Service whose radio network was invaluable in such a remote area. I particularly thank my co-worker at JSN, Mike Smith; useful advice and discussion on ideas in this paper were provided by him, Bob Wasson and Barry Cundy. I am also especially grateful to Win Mumford for her work on the maps and illustrations. References BURKE, R. O. & WILLS, W. J. 1861. 'The Burke and Wills Exploring Expedition'. Wilson & Mackinnon: Melbourne. CALLEN, R. A. and BRADFORD, J. 1992. The Cooper Creek fan and Strzelecki Creek - hypsometric data, Holocene sedimentation and implications for human activity. Mines and Energy Review, South Australia 158: 52-57. HISCOCK, P. 1981. Raw material rationing as an explanation of assemblage differences: a case study of Lawn Hill, north-western Queensland. Pp. 178- 190 in 'Archaeology at ANZAAS, Canberra'. Second printing. Ed. G K Ward. Canberra Archaeological Society: Canberra. HUGHES, P. J. & LAMPERT, R. J. 1980. Pleistocene occupation of the arid zone of south- east Australia: research prospects for the CoopeT Creek-Strzelecki Desert region. Australian Archaeology 10: 52-67. JONES, W. 1979. 'Up the creek: hunter-gatherers in the Cooper Basin.' B.A. (Hons) thesis, University of New England: Armidale. KERWIN, B. & BREEN, J. G. 1981. The land of stone chips. Oceania 51: 236-31 1. ARCHAEOLOGY OF THE MIDDLE COOPER BASIN 91 KREIG, G. W. & CALLEN, R A. 1990. Coongie Lakes Control Zone - a map and note for the surface geology and terrain. Mines and Energy Review, South Australia 157: 59-61. LAYTON, R, FOLEY, R, & WILLIAMS, E. 1991. The transition between hunting-and-gathering and specialized husbandry of resources. Current Anthropology 32: 255-274. MCBRYDE, I. 1987. Goods from another country: exchange networks and the people of the Lake Eyre Basin. Pp. 252-273 in 'Australians to 1788'. Ed. D. J. Mulvaney & J. P. White. Fairfax, Syme and Weldon Associates: Sydney. MCKINLAY, J. 1862. 'McKinlay's Journal of Exploration in the Interior of Australia'. F. Balliere: Melbourne. MACUMBER, P J. 1980. The influence of groundwater discharge on the Mallee landscape. Pp. 67-84 in 'Aeolian Landscapes in the Semi-arid Zone of South- eastern Australia. Eds. R. R. Storrier & M E Stannard. Australian Society of Soil Science: Wagga Wagga. MEEHAN, B. 1982. 'Shelibed to Shell Midden'. Australian Institute of Aboriginal and Torres Strait Islander Studies: Canberra. REID, J. R. W. & GILLEN, J. 1988. The Coongie Lakes Study. Unpublished consultancy report to the South Australian Department of Environment and Planning: Adelaide. REID, J. R. W. & J. T. PUCKERIDGE. 1990. Coongie Lakes. Pp. 119-131 in 'Natural history of the north- east deserts'. Ed. M. J. Tyler, C .R. Twidale, M. Davies and Wells, C. B. Royal Society of South Australia Occasional Publications 5. Royal Society of South Australia: Adelaide. REUTHER, J. G. 1981. 'The Diari'. Volumes 1-13. Translated by the Reverend P. A. Schcrer. Australian Institute of Aboriginal and Torres Strait Islander Studies, Canberra. SMITH, M. A. 1986. The antiquity of seed-grinding in Central Australia. Archaeology in Oceania 21: 29-39. SMITH, M. A., WILLIAMS, E, & WASSON, R. J. 1991. The archaeology of the JSN site: some implications for the dynamics of human occupation in the Strzelccki desert during the late Pleistocene. Records of the South Australian Museum 25: 175- 192. TWIDALE, C. R. 1972. Landform development in the Lake Eyre region, Australia. The Geographical Review 62: 40-70. VETH, P. M. 1993. Islands in the Interior: the dynamics of prehistoric adaptations within the arid zone of Australia. International Monographs in Prehistory. Archaeological Series 3 VETH, P. M, HAMM, G. & LAMPERT, R. J. 1990. The archaeological significance of the Lower Cooper Creek. Records of the South Australian Museum 24: 43-66. WASSON, R J. 1986. Geomorphology and Quaternary history of the Australian continental dunefields. Pp. 419-432 in J. C. Vogel, Ed. 'Late Cainozoic Palaeoclimates of the Southern Hemisphere'. A. A. Balkema: Rotterdam. WILLIAMS, E. 1988. The archaeology of the Cooper Basin: report on fieldwork. Records of the South Australian Museum 22: 53-62. REVISION OF AUSTRALIAN AMPHIOPS ERICHSON, ALLOCOTOCERUS KRAATZ AND REGIMBARTIA ZAITZEV (COLEOPTERA: HYDROPHILIDAE) C.H.S. WATTS Summary The Australian members of the Hydrophilid genera Amphiops (five species), Allocotocerus (three species) and Regimbartia (one species) are revised and redescribed. Keys to species of Amphiops and Allocotocerus are given. Amphiops austrinus, Amphiops micropunctatus and Allocotocerus yalumbaboothbyi are described as new. REVISION OF AUSTRALIAN AMPHIOPS ERICHSON, ALLOCOTOCERUS KRAATZ AND REGIMBARTIA ZAITZEV (COLEOPTERA: HYDROPHILIDAE) C. H. S. WATTS WATTS, C. H. S. 1998. Revision of Australian Amphiops Erichson, Allocotocerus Kraatz and Regimbartia Zaitzev (Coleoptera: Hydrophilidae). Records of the South Australian Museum 30(2): 93-106. The Australian members of the Hydrophilid genera Amphiops (five species), Allocotocerus (three species) and Regimbartia (one species) are revised and redescribed. Keys to species of Amphiops and Allocotocerus are given. Amphiops austrinus, Amphiops micropunctatus and Allocotocerus yalumbaboothbyi are described as new. C. H. S. Watts, South Australian Museum, North Terrace, Adelaide, South Australia 5000. Manuscript received 15 April 1997. The three genera included in this revision are grouped for convenience not phylogeny although both Allocotocerus Kraatz, 1883 and Regimbartia Zaitzev, 1908 belong in the same tribe (Berosini) and they and Amphiops Erichson, 1843 are frequently mixed together in collections. All are wet-tropical beetles with, at the generic level, a wide distribution outside Australia (Hansen 1991). They are all relatively small very deep- bodied insects with Amphiops and Allocotocerus having almost spherical bodies. They occur commonly in pools and swamps around the tropical coast from the Kimberley to about Sydney. The most recent taxonomic work on the Australian species was by J. Balfour- Browne who briefly commented on them and described Allocotocerus tibialis and Amphiops queenslandicus (Balfour-Browne 1939) and Hansen, 1991 who discussed their generic placement. Material was examined from the following collections. AM Australian Museum, Sydney ANIC Australian National Insect Collection BM(NH) Natural History Museum, London NMV Museum of Victoria NTM Northern Territory Museum QDPIM Queensland Department of Primary Industries, Mareeba QM Queensland Museum, Brisbane SAMA South Australian Museum, Adelaide UQIC University of Queensland Insect Collection, Brisbane WAM Western Australian Museum, Perth. Systematics The three genera can be separated from other Australian aquatic Hydrophilids by the following characters (after Hansen 1991). Amphiops: Size 2.5^.5 mm. Dark brown to black. Eyes divided into upper and lower portions by extensions of side of head. Elytra approximately as high as long, without striae. Meso- and meta-tibiae without swimming hairs. First ventrite very short, with a fringe of fine setae rising from its basal margin. Allocotocerus: Size 3.5^.5 mm. Black. Meso- and meta-tibiae with swimming hairs. Eyes normal. Elytra as high as long, virtually without striae. Regimbartia: Size 3.5-5.0 mm. Black. Meso- and meta-tibiae with swimming hairs. Eyes normal. Elytra high, about 2.8x longer than height, with distinct striae. For more detailed descriptions and discussion of the affinities of these genera see Hansen, 1991. Amphiops Erichson, 1843 Australian Amphiops are all very similar and are best separated by the male genitalia. Indeed I have found it impossible to reliably separate A. queenslandicus J. Balfour-Browne and A. duplopunctatus Blackburn by any other means. The best general character is the form of the clytral punctation, in particular the interstrial punctation on the sides of the elytra. There are three more or less distinct size classes of elytral interstrial punctures: 1) large, about the size of the strial punctures; 2) small, between about 20- 94 C. H. S. WATTS 60% of the larger ones; 3) micro, which are normally no more than pin pricks even under moderate magnification. Key to Australian Amphiops Interstrial punctures consisting of a few large punctures and a few to numerous micro punctures, the small size seemingly absent (Figs 1&3). Systematic punctures on pronotum large and distinct 2 Interstrial punctures consisting of large, small and micro punctures with small predominating (Figs 2 & 4) 3 Central lobe of aedeagus not hooked (Fig. 13). First elytral stria (close to suture) distinct, traceable (anteriorly) well beyond apex of scutellum. Scutellum always moderately punctate A. micropunctatus sp.nov. Central lobe of aedeagus hooked at tip (Fig. 14). First elytral stria virtually absent. Scutellum lacking punctures or weakly punctate with very small punctures A. australicus Blackburn. Central lobe of aedeagus hooked at tip (Fig. 9). Small punctures on sides of elytra relatively small, < 1/4 diameter of adjacent large punctures (Fig. 4). Serial punctures easily traceable at apex A. austrinus sp.nov. Central lobe of aedeagus rounded, parameres straight. Small punctures on sides of elytra relatively large, many 1/2 size of adjacent punctures (Fig. 2). Serial punctures may be hard to trace at apex 4 Basal portion of aedeagus shorter than parameres (Fig. 10). Size <3.9 mm. Elytra often with a vague series of darker patches in alternate interstriae A. duplopunctatus Blackburn Basal portion of aedeagus nearly twice length of parameres (Fig. 11). Size > 3.5 mm. Elytra without series of dark patches A. queenslandicus J. Balfour-Browne. Amphiops austrinus sp. nov. Description (number examined 5) Figs 4, 9 Length 3.4-4.0 mm. Broadly oval, elytra deep, almost as high as long. Reddish-brown, disk of pronotum a little darker as are serial punctures. Head broad, strongly punctate with variously sized punctures, confluent at front. Pronotum broad, punctures of various sizes, weak on disk grading to strong laterally, systematic series distinct. Elytron very weakly punctured near suture, grading to strong laterally; in middle at side serial punctures strong, large, somewhat difficult to trace; large punctures separated by own diameter or a bit less, small punctures, which alternate with larger ones, are very much smaller; interstrial punctures of three distinct sizes, largest as large or larger than those in serial lines, separated by own width or less, small punctures more numerous, very much smaller, 1/4-1/3 diameter of large ones, in places third size group of minute (micro) punctures present. Aedeagus squat, basal portion shorter than parameres. Apical portion of central lobe narrow, weakly tapering, considerably shorter than parameres, ending in small broad hook. Parameres broad in basal half, narrow in apical, strongly curved downwards near tip, tips truncated and a little enlarged. Distribution Only known for the type localities in south-east Queensland. Types Holotype: 'Brisbane 1/64, CW\ SAMA. Paratypes: 4, 'Qld Pctrie, 10 km W, 23/1 1/95, C. Watts', SAMA. Remarks In most characters similar to A. duplopunctatus and A. queenslandicus. Apart from the male genitalia A. austrinus differs from both of these by the greater contrast in size between the large and small interstrial punctures on elytra, caused primarily by the comparatively small size of the small punctures. In A. duplopunctatus and A. queenslandicus the small interstrial punctures become quite prominent towards the side and apex, often reaching half the size of the larger ones. In A. austrinus, apart from very near the elytron edge, the small punctures remain much smaller than the larger ones. The downward curve of the parameres and the apical hook on the tip of the central lobe of the aedeagus, also clearly set it apart from these two species. Amphiops australicus Blackburn, 1898 Description (number examined 234 ) Figs 3, 14 Length 2.7^.0 mm. Broadly oval, elytra deep, about as high as long. Brown, sometimes with distinct darker blotches on elytron, to black with AMPHIOPS, ALLOCOTOCERUS AND REGIMBARTIA 95 £&&&#$■ Oo oa : o : ^o^ 6 o'Sf^° o : , O'-O ' ^°^ o. o. 0-.9 ■■-. n Y5V 3.5 mm). Balfour-Browne (1939), considered A. duplopunctatus to be possibly no more than a subspecies of A. australicus, an opinion based mainly on a perceived close similarity of the male genitalia. The male genitalia of what I take to be A. australicus and A. duplopunctatus are clearly distinct (the types are female). The markedly different punctation also differentiates the two species. Amphiops micropunctatus n. sp. Description (number examined 30) Figs 1,13 Length 2.7-4.3 mm. Oval, elytron deep, a little longer than high. Reddish-brown, punctures on elytron darker. Head broad, moderately strongly punctate, punctures tending to be either big or small, smaller predominating, becoming confluent on front margin. Pronotum broad, moderately punctate, punctures of varying sizes, stronger laterally, systematic series distinct, comparatively large, many over twice size of adjacent punctures, nearly twice size of eye facet. Elytra weakly and shallowly punctured on disc, becoming much stronger towards sides. Laterally in middle serial punctures large, easy to trace, about a puncture-width apart; interstrial area shiny, punctures of two sizes, the larger sparse, most > 2x width apart, somewhat smaller than strial, the smaller, more numerous, very small (micro), in contrast to other punctures, becoming obsolete laterally. Aedeagus with basal piece about same length as parameres, parameres wide in basal half, narrow in apical, tips broadly rounded, central lobe narrow, tip bluntly pointed, reaching nearly to end of parameres. Remarks In colour and size resembling A. duplopunctatus but differing from it in punctation and in the shape of the aedeagus which has the central lobe larger than in A. duplopunctatus. The absence of the 'small' size of interstrial punctures in the species resembles A. australicus and readily separates these two species from the other Australian Amphiops which have the interstrial areas much more punctate. From A. australicus (and most other Australian Amphiops) it differs by having the first (sutural) elytral stria quite well developed as well as a greater development of very small (micro) punctures on the elytra, although an occasional A. australicus may have quite extensive micro punctation. This species is only known from a limited area of North Queensland. Judging from the localities it is possible that it is a closed forest species. The specimens from the West Claudie basin and Lockhart River were taken in October from still pools in closed forest. No other habitat information is available. Types Holotype male: 'Australia, N Qld 15 km WNW of South Johnstone "Light Trap' 11 , 'Fay and Halfpapp 1986', QDPIM. ** Paratypes: 2 same data as holotype, QDPIM; I, 'Australia N Qld Tolga 3', '1986 J. D. Brown light Trap', QDPIM; 1, 'Cow Bay, N of Daintree AMPHIOPS, ALLOCOTOCERUS AND REGIMBARTIA 99 11 13 14 FIGURES 9-14. Ventral view of aedeagus. 9, Amphiops austrinus; 10, A. duplopunctatus; 11, A. queenslandicus\ 12, Regimbartia attenuata\ 13, Amphiops micropunctatus; 14, /*. australicus. 100 C. H. S. WATTS N Qld, 15-22nd March 1984 I. C. Cunningham', QDPIM; 2, '12°43S 143°17E, 9 km ENE of Mt Tozer 5- 10th July 1986 T Weir and A. Calder\ ANIC; 3, '12.44S 143. 14E 3 km ENE of Mt Tozer 28th Jun - 4th July 1986 T Weir and A. Calder', ANIC; 10, '15°03S 145°09E 3 km NE Mt Webb 30th April - 3rd May 1981 A Calder', 8 ANIC, 2 SAMA. Distribution Queensland Dead Horse Ck, ANIC; East Claudic R, UQIC; Gordon Ck nr Claudie R, UQIC; Iron Range, UQIC; 9km NNW Lockhart R, ANIC; 3.5km SW by S Mt Baird, ANIC; 2km NNE of Mt Tozer, ANIC; 6km ENE Mt Tozer, ANIC; 11km ENE Mt Tozer, ANIC; Mt Webb Nat. Pk, ANIC; West Claudie R, ANIC. Amphiops queenslandicus Balfour-Browne, 1939 Description (number examined 34 ) Fig. 1 1 Length 3.5-5.3 mm. Broadly oval, elytra deep, almost as high as long. Dark reddish-brown, elytra with vague darker/lighter areas and many serial punctures black. Head broad, strongly punctate, punctures of varying sizes becoming confluent along front margin. Pronotum broad, moderately punctate on disc becoming stronger toward sides, systematic punctures moderate, about 3x size of adjacent punctures, a little larger than eye facet. Elytra with small but sharp punctures on disc becoming much stronger laterally. At sides in middle, large serial punctures relatively small, most separated by more than their widths, small alternating punctures relatively large, some approaching 1/2 width of larger ones; interstrial punctures of three sizes: large, larger than serial punctures, separated by more than their width, small much more numerous, relatively large, 1/4- 1/2 diameter of large ones, micro punctures present over most of elytra. Aedeagus elongate, basal piece nearly twice length of parameres. Central lobe shorter than parameres, narrowing to a point, apex not swollen or hooked. Parameres broad basally, becoming narrower apically, tips rounded, weakly curved downwards. handwritten locality (which I cannot read), BM(NH). Seen. Paratypes: 6, same data as Holotype (except handwritten label and Type designation); 1, male, 'Rockingham austral', 'Australia' Amphiops queenslandicus M. J. Balfour-Browne det., all in BM(NH) with round paratype labels. Seen. Distribution Northern Territory S Alligator R, QM; Black Point Coburg Pen, ANIC; 5km NNW Cahills Crossing, ANIC; 7km NW by N Cahills Crossing, ANIC; Coastal Plains Research Station, ANIC; Daly R Mission, ANIC; Darwin, SAMA; Fogg Dam, NTM; Howard Springs, SAMA; Humpty Doo, ANIC; Jabiru, NTM; Jim Jim, ANIC, NTM; Kakadu NP, NTM; Kapalga, QM; Koongarra, ANIC; 12km NNW Mt Cahill, ANIC; 19km WSW Mt Cahill, ANIC; 9km N by E of Mudginbarry HS, ANIC; Oenpelli, AM; 6km SW by S Oenpelli, ANIC; 12°40S 132°22E, QM. Queensland Ashgrove, QM; Ayr UQIC; Bowen, SAMA; Brisbane, UQIC; Bulburin St Forest via Many Peaks, UQIC; Caloundra, SAMA; Cooloola, QM; East Claudie R, UQIC; Gatton, UQIC; Homehill, SAMA; Ingham, UQIC; Ipswich, AM; Iron Range, UQIC; Lockerbie UQIC; Many Peaks, UQIC; S Pine R, QM; Ripley, SAMA; Rockhampton, ANIC, SAMA; 20 km S Townsville, SAMA; 37 km S Townsville, SAMA. Remarks The distinctive male genitalia readily identify this species. Amphiops duplopunctatus is smaller with the elytra not quite as deep, elytral punctures weaker and usually differently coloured (see under A. duplopunctatus). Amphiops queenslandicus also closely resembles A. austrinus but in this case the weaker, smaller punctures on the elytra of A, austrinus seem different enough to enable reliable separation. However, the male genitalia should be used wherever possible. Regimbartia Zaitzev, 1908 Types Holotype male: 'Queensland Australia', 'Amphiops queenslandicus Type J. Balfour- Browne det.', with round Type label and a Regimbartia attenuates (Fabricius,1801) = Volvulus scaphiformis Fairmaire, 1879 Synonymy after d'Orchymont, 1932. AMPH10PS. ALLOCOTOCERUS AND REGIMBARTIA 101 Description (number examined 89) Fig. 12 Length 3.6-5.0 mm. Narrowly boat-shaped, high, sides of elytra subparallel, perpendicular, black, shiny, appendages testaceous. Head deflexed, eyes large, punctures moderate scattered, weaker on disc, those forward of eye about size of eye facet, systematic punctures inward from eye relatively small, about size of eye facet. Pronotum broad, weakly and sparsely punctured on disc, more strongly towards sides where they are about same size as on front of head; systematic punctures few and hard to trace. Elytra more strongly punctured, somewhat stronger at sides than on disc, each elytron with 10 serial lines, inner two incomplete anteriorly, for the most part serial punctures joined forming sharp grooves. Mesosternum with strong narrow keel, apical sternite with short sharp spine in midline at apex. Male: Protarsi with two basal segments moderately expanded dorso-vcntrally. Types Hydrophilus attenuate! Fabricius. Ceylon. Type not located. Volvulus scaphiformis ? Holotype. Rockhampton Fairmaire, ?in Museum Godeffroy. Not seen. Distribution Northern Territory Bessie Springs, ANIC; 11 km SW by S of Borroloola, ANIC; 30 km NE by E of Borroloola, ANIC; 48 km SW by S Borroloola, ANIC; 1 km N of Cahills Crossing, ANIC; 7 km NW by N Cahills Crossing, ANIC; Daly River, NTM Darwin, ANIC, SAMA; Elsey Creek, ANIC Fcrgusson R, ANIC; Fogg Dam, ANIC, NTM Howard Springs, SAMA; Jabiru, NTM Katherine, NTM; Keep River NP, ANIC; Lake Bennett, NTM; Magela Ck, ANIC; Manton Reservoir, NTM; 2 km N Mudginberri HS, ANIC; 6 km N by E Mudginberri, ANIC; Nourlangie Ck, ANIC; Roper R, ANIC; South Alligator R, QM; U.D.P. Falls, NTM; Queensland Archer Bend, SAMA; Ayr, UQIC; Brisbane. MV, QM, UQIC; Cairns, ANIC, QM; Calliope R, ANIC; Caloundra, SAMA; Cape Flattery, DPIM; Chillagoe, DPIM; Claudie R, UQIC; East Claudie R, QIC; Clermont, AM; Cooktown, MV; 40M N Cooktown, UQIC; Duaringa, AM; Edungalba, ANIC; Einasleigh, DPIM; Frenchman's Creek via Rockhampton, UQIC; Gin Gin, UQIC; Gregory River Hotel, DPIM; Homehill, SAMA; Kirrama Rng, QM; Karumba, UQIC; Lakefield NP, ANIC; Mackay, SAMA; Mareeba, ANIC, DPIM; 18km N Mareeba, DPIM; Mary Ck, ANIC; Mclvor R, UQIC; Mt Garnet, SAMA; 3.2 km SW of Mt Inkerman, ANIC; Mt Molloy, ANIC, QM; 3 km ENE Mt Tozer, ANIC; 6 km ENE Mt Tozer, ANIC; Nordcllo Lagoon, DPIM; 11km WSW Petford, DPIM; N Pine R, QM, UQIC; Rockhampton, SAMA; 67 km E Roma, QM; Samford, UQIC; Silver Plains CP, DPIM; 15km WNW South Johnstone, DPIM; Starbright HS, DPIM; Strathmore Stn, DPIM; Tolga, DPIM; Townsville, QM; Walkamin, DPIM; 40km S Weipa, DPIM; Yaamba, UQIC; Yeppoon, UQIC; 18°38 S 138°11 E, ANIC; 18°34 S 138°08 E, ANIC. Western Australia Channley R, ANIC; Derby, SAMA; Fitzroy R, ANIC; 12 km S Kalumburu Mission, ANIC; Kunanurra, DPIM; 1 km NNE Millstream, ANIC; Synnot Ck, ANIC; 14°53 S 125°45 E, SAMA. New South Wales Eccleston, UQIC. Remarks Once known, this distinctive water beetle is unlikely to be confused with any other Australian species. It is relatively common in shallow swamps and dams along the coast from northern New South Wales to the Kimberlcy. Beyond Australia the species has a wide distribution in South-east Asia as far west as Sri Lanka. I have followed d'Orchymont (1932 p. 709) for the name of this widespread species. Allocotocerus Kraatz, 1883 The three Australian species of Allocotocerus (- Globaria Latreille) are all moderately sized, highly spherical insects with the ventral surface shiny black. Within Australia they can only be confused with species of Amphiops, particularly A. queens landicas which is of a similar size and shape. The even, regular punctation, uniformly black ventral surface, and swimming hairs on the legs in Allocotocerus readily separate them from Amphiops. There is little to separate the species other than the sexual characters of the aedeagus, trochanter setae and the labrum. The first described and best known species, A, punctatus, is common and widespread in slowly 102 C. H. S. WATTS moving water and in swamps in eastern Australia bit larger than size of eye facet, systematic from northern New South Wales northwards. It punctures virtually absent. Pronotum narrower appears to be absent from the Northern Territory than elytra, punctures on disc much weaker than and northern Western Australia where its place is on head, becoming considerably stronger towards taken by A. tibialis and A. yalumbaboothbyi. front comers, systematic punctures few, sparse Conversely these species likewise have not been and hard to find. Elytra covered with very even recorded from the east coast. punctures in same area but much stronger laterally An additional species (the type species, A. than on disc, serial punctures 1.5-2x diameter of bedeli Kraatz 1883) occurs in New Guinea, adjacent punctures, reduced to 2-3 short lines Unfortunately the holotype does not appear to be midway along side of elytron. Midline of in the Deutsches Entomologisches Institut in mesosternum with strong, tall spike between Eberswalde. In its place is a label written by mesocoxae. First ventrite constricted in middle Korschefsky in 1937 indicating it was lost (Lothar with rather wide strong midline carina. Zerche pers. com.). The recognition of this species Metasternum with two strong carinae in midline and any possible influence on the nomenclature with deep gap between them, the anterior one of Australian species must await the collection of cylindrical and angled backwards, the posterior further specimens from New Guinea. one anvil shaped in lateral view. Male: Labrum prominent, front edge deeply bifid. Protibia triangular in cross-section. Key to Australian Allocotocerus Aedeagus with basal portion strongly bent, 2.5x length of parameres which are broad for whole Labrum strongly extended forward, strongly | engthj central lobe broad e t for short blfld male ^ P mctatUS (B^kburn) narrow port]0n at apcx MeSQ _ and ^ Labrum normal, often hidden beneath clypeus, trochanters with widely scattered short, stout setae not blfid - 2 on ventral face. Protibia light yellow with darker apical portion Female: Labrum little exposed, front edge not A.punctatus (Blackburn) bifid. Protibia of normal shape, contrast between Protibia lighter than other legs but without dark a P ical P ortion and yellow on rest of proleg darker apical portion 3 often less pronounced than in male. Meso- and meta-trochanters with scattered, sparse, short, Setae on outer posterior angle of ventral stout setae on ventral face. surface ot mesofemur more developed than those on metafemur (Figs 7 & 8). Aedeagus Type with parameres broad with rounded tips, \j , . n^n«oi *™ », ™ *^*^ .***,>, central [lobe sharply pointed, much shorter than Holotype: T2328' 'Blackburn coll 1910-236', parameres (Fig. 16) Volutus punctatus, Blackburn', BM(NH). Seen. A. tibialis (Balfour-Browne) _ . „ c ., . Distribution Setae on metafemur equally poorly or slightly more developed than on mesofemur (Figs 5 & Queensland 6). Aedeagus with parameres narrowing to 8 km N Bluewater, SAMA; Cardstone, ANIC point, central lobe broad with narrow spine at 40 km N Coen, SAMA; 70 km N Coen SAMA" tip, as long as parameres (Fig. 17) Caloundra, SAMA; 29 km NW by W Cooktowni A -y^rnbaboothbyi, V .no,. ANJC; 4Q km N Cooktown? SAMA; , ngham ANIC; Lakefield NP, QDPIM; 18 km N Mareeba, QDPIM, ANIC; Mary Ck, ANIC; Mc Ivor R, ANIC; Townsville, MV; Moorehead R, ANIC; Moreton, ANIC; Mt Coolum, ANIC; 20 km S Townsville, SAMA; 37 km S Townsville, SAMA; Weniock R Crossing, ANIC; 15°17S 145°10E, ANIC. 1 — 2 — Allocotocerus punctatus (Blackburn, 1888) Description (number examined 112) Fig. 15 Length 3.5-4.5 mm. Oval, elytra high, height only a little less than length. Black, shiny, appendages lighter, yellowish with darker terminal portion to tibia. Head narrow, front margin broadly and quite deeply concave, sides of clypeus narrowly margined, rather evenly covered with moderate punctures which are weaker in middle than at sides, those forward from eyes a Remarks In overall shape, punctation and colour closely similar to A. tibialis and A. yalumbaboothbyi but the male characters readily separate it from these species. Male A. yalumbaboothbyi and A. tibialis AMPHIOPS, ALLOCOTOCERUS AND REG1MBARTIA 103 15 16 17 FIGURES 15-17. Ventral view of aedeagus. Allocotocerus valumbaboothbvi. 15, Allocotocerus punctatus; 16, Allocotocerus tibialis; 17. lack the very distinctive labrum and protibia of male A. punctatus. In A. punctatus the dark apical portion of the protibia separates both male and female from the other species which have more uniformly coloured protibiae. In both male and female A. tibialis, the number and density of setae on the trochanters are much larger than in A. punctatus. The species is relatively common in coastal dams and swamps from northern Southern Queensland to Cape York, often found together with the similarly shaped Amphiops. Allocotocerus tibialis (Balfour-Browne, 1939) Description (number examined 61) Figs 7, 8, 16 Length 4.5-5.0 mm. Oval, elytra high, height 104 C. H. S. WATTS only a little less than length. Black, shiny, appendages dark-testaceous. Head narrow, front edge widely and moderately concave, sides of clypeus margined, strongly punctured, punctures in front of eyes larger than eye facet, systematic punctures inwards from eye few and hard to trace, punctures on disc somewhat weaker than elsewhere, about size of eye facet. Pronotum narrow, somewhat more weakly punctured than head particularly on disc, systematic punctures few and hard to find. Elytron punctured rather more strongly than on pronotum and head, punctures much weaker on disc than on sides where they are strong and less than a puncture-width apart. Serial punctures few or absent, if present little larger than adjacent punctures. Mesosternum with moderately tall, sharp spine in midline. Midline of metasternum with two distinct carinae, separated by deep gap, anterior carina cylindrical and angled strongly backwards, posterior one broader and curving backwards and upwards when viewed from side. First abdominal ventrite narrowed in middle half, midline with rather narrow and strong carina, ventrites two and three narrowed in central half, weakly bulbous/carinate in midline. Area between ventrites deeply grooved. Fourth ventrite weakly narrowed anteriorly, smooth, Male: Labrum simple. Protibia a little flattened and widened, tarsi unmodified. Mesotrochanters stout with a well-marked group of long golden setae on anterior apical ventral angle, inner half of ventral face lacking punctures/setae. Metatrochanters stout with a less well-developed group of setae in same place, metatibia stout. Aedeagus long, sinuate, relatively broad, basal piece about 1.5x length of parameres, parameres broad, subparallel, rounded at tips, central lobe narrowing progressively to tip, shorter than parameres. Types Holotype male; 'Adelaide River 92-2\ '4969', 'Globaria tibiale mihi J. Balfour-Browne det.' with red type label, BM(NH). Seen. Lectotypes: 1, 'Adelaide River, NW Australia J. J. Walker', L G. C. Champion Coll. B,M. 1927- 409' with red (allotype) type label, BM(NH), 10 same data, in BM(NH); 3, same data as holotype, BM(NH). Seen. Distribution Northern Territory Gimbat Sin, NTM; 6 km E Humpty Doo, QDPIM; Jabiru, NTM; 10 km N Jabiru, QDPIM; Junction of Arnhem Hwy and Oenpelli Rd, NTM; Magela Ck, ANIC; 19 km E by S Mt Borradaile, ANIC; 8 km E Mt Cahill, ANIC; 46 km WSW Mt Cahill, SAMA; Nabarlek Dam, ANIC; Nourlangie, ANIC; UDP Falls, NTM; Wildman R,NTM. Remarks Both male and female A. tibialis can be separated from A. punctatus and A yalumbaboothbyi by the more extensive development of setae on the mesotrochanters. The elytral striae are also usually much weaker or absent in this species but at least traceable in the others. The species has not been recorded from the east coast and appears to have a rather restricted distribution in coastal Northern Territory. Here it occurs commonly, often with A. yalumbaboothbyi. Amphiops tibialis can also be separated from the very similar A. yalumbaboothbyi by the differently shaped posterior mesosternal protuberance: in A. tibialis this is thicker and curved upwards behind in lateral view whereas it is straight in A. yalumbaboothbyi. Allocotocerus yalumbaboothbyi sp. nov. Description (number examined 51) Figs 5, 6, 17 Length 3.5^.5 mm. Oval, elytra high, only a little longer than high. Black, shiny, ventral surface and appendages dark-testaceous, proleg yellowish. Head narrow, front margin broadly and quite deeply concave, rather evenly covered with moderate punctures which are a little weaker on disc than at sides, those forward from eyes a bit larger than size of eye facet, systematic punctures virtually absent. Pronotum narrower than elytra, punctures on disc much weaker than on head, becoming considerably stronger towards front corners, systematic punctures few and hard to find. Elytra covered with very even punctures which become much stronger laterally, serial punctures reduced to 2-3 short lines on each elytron towards sides in middle; midline of mesosternum with tall sharp spike forward of mesotrochanters, midline of metasternum with two strongly raised carinae separated by short deep gap, front carina in shape of backwardly inclined broad spine, rear carina oval shaped on a broad pedestal when viewed laterally. First ventrite moderately constricted in middle with a narrow but well marked central carina extending AMPHIOPS, ALLOCOTOCERUS AND REGIMBARTIA 105 forward between metatrochanters. Second sternite also constricted but without carina, fourth not restricted. Sutures between ventrites wide, deep, well-marked. Male: Labrum small, usually contracted, not bifid, protibia not expanded, basal two segments of protarsi a little expanded with ventral hairs. Mesotrochanters with a few scattered, short spines on apical ventral surface. Mesotrochanters with longer more numerous setae forming a sparse brush along posterior ventral portion of metatrochanter. Aedeagus long, sinuate with relatively broad basal portion which is about 2.5x length of parameres. Parameres narrow, evenly narrowing to quite sharp tip. Central lobe as long as parameres. In a number of specimens the parameres are twisted and splayed out, presumably an artefact of preservation, but one that I have not seen in other species. Female: Basal segment of protarsi not weakly expanded, with ventral hairs. Mesotrochanters with a few sparse, stout setae on ventral face at apex, metatrochanter with similar setae restricted to apical half. Types Holotype male: '12°22S 133°01E 6 km SW by 5 of Oenpelli NT, 30. v. 73, at light E. G. Matthews', SAM A. Paratypes: 4, 'W Australia Mitchell Plateau 14°40S 125°44E 23 Sept 1982 B. V. Timms\ SAMA; 4, '13°34S 132°15E\ 'Moline Rockhole' 6 km E by E of Mt Daniela 23. v. 1974, NT T. Weir and T. Angeles', NTM; 6, '12°32S 132°50E Koorgarra 15 km E of Mt Cahill NT, 15.xi.1972, T. Weir and A. Allwood', NTM; 11, 'NT UDP Falls, 18-19th July 1980, M. B. Malipatil Freshwater Pool', NTM; 20, '12°49S 132°51E, 15 km E by N of Mt Cahill NT, 29.X.72, light trap, E. Britton 1 , ANIC. Distribution Northern Territory 8 km E by N of Mt Cahill, NTM; 19 km E by N Mt Cahill, NTM; Junction of Amhem Hwy and Oenpelli Rd, NTM; Kambolgie Ck, SAMA; Koongarra, NTM; Wildman R, NTM. Remarks Virtually indistinguishable from A. tibialis and A. punctatus other than by sexual characters. Males can be easily separated from A. punctatus by the lack of a large bifid labrum, from A. tibialis by the weaker development of setae on the mesotrochanters. In A. tibialis the mesotrochanter setae are longer and denser often to the extent of matting together. They are also more restricted to the apical portion of both meso- and meta- trochanters than in A. yalumbaboothbyi. Separation of female specimens is more difficult. Female A, punctatus have the prolegs pale yellow with a more or less obviously darker terminal portion to the tibia. In both A. yalumbaboothbyi and A. tibialis the tibiae lack the darker apical portion. The setae on both meso- and meta-trochanters of female A. yalumbaboothbyi and A. punctatus are small and sparse. In A. tibialis they are more developed particularly on the mesotrochanter. Acknowledgments The curators of the collections listed earlier provided ready access to specimens in their care. Ms D. Churches and Ms C. Home typed the manuscript and Mr R. Gutteridge prepared the drawings. Mrs M. Anthony and Mrs J. Evans helped with library aspects. Dr E. Matthews greatly improved the manuscript. I thank all these people without whom my task would have been much greater. References BALFOUR-BROWNE, J. 1939. Contribution to the study of the Palpicornia Part III. Annals and Magazine of Natural History 11(4): 289-310. BLACKBURN, T. 1888. Notes on Australian Coleoptera with descriptions of new species. Proceedings of the Linnean Society of New South Wales 111(2): 805-875. BLACKBURN, T. 1898. Further notes on Australian Coleoptera, with descriptions of new genera and species. Transactions of the Roval Society of South Australia 22: 221-233. ERICHSON, W. F. 1843. Beitrag zur Insecten-Fauna von Angola, in: besondere Beziehung zur geographischen Verbreitung der Insecten in Afrika. Archiv fur Naturgeschichte 9: 199-267. FABRICIUS, J. C. 1801. Systema Eleutheratorum...I, 24+506pp. Kiliae. FAIRMAIRE, L. 1879. Descriptions dc Coleopteres nouveaux ou peu connus du Musee Godeffroy (Hydrophilidae). Journal Museum Godeffroy 24: 80-83. 06 C. H. S. WATTS HANSEN, M. 1991. The Hydrophiloid Beetles. Phylogeny, Classification and a Revision of the Genera (Coleoptera: Hydrophiloidea). Biologiske Shifter 40: 367pp. KRAATZ, G. 1883. Allocotocerus nov. gen. Hydrophilidarum. Deutsche Entomologische Zeitschnft 27: 14-15. ORCHYMONT, A. 1932. Zur Kenntnis der kolbenwasser Kafer (Palpicomia) von Sumatra, Java und Bali. Archives Hydrobiologica. Supplement Band IX (Tropische Binnengewassen 1 1): 623-714. ZAITZEV, F. A. 1908. Catalogue des Coleopteres aquatiques des families Dryopidae, Georyssidae, Cyathoceridae, Heteroceridae et Hydrophilidae. Trudy russkogo Entomologicheskogo Obshchestva 38: 283-420. DOUGLAS MAWSON: THE GEOLOGIST AS EXPLORER D. W. CORBETT Summary Douglas Mawson was a born explorer. His professional career as a geologist and academic in Australia demonstrates this aspect of his character as clearly as his exploits in the Antarctic. Science and exploration dominated his life and motivated his endeavours, and the inseparable link between them was his pursuit of knowledge of glaciation-both ancient and modem. This paper reviews the diversity and wide geographical extent of his geological investigations, principally in South Australia, during the first half of the twentieth century at a time when the geology of large areas of the State was still largely unknown. The manner of his achievement and his distinguished contribution to the advance of geological knowledge in South Australia exemplify Mawson as the explorer-geologist. DOUGLAS MAWSON: THE GEOLOGIST AS EXPLORER D. W. CORBETT CORBETT, D. W. 1998. Douglas Mawson: the geologist as explorer. Records of the South Australian Museum 30(2): 107-136. Douglas Mawson was a born explorer. His professional career as a geologist and academic in Australia demonstrates this aspect of his character as clearly as his exploits in the Antarctic. Science and exploration dominated his life and motivated his endeavours, and the inseparable link between them was his pursuit of knowledge of glaciation — both ancient and modern. This paper reviews the diversity and wide geographical extent of his geological investigations, principally in South Australia, during the first half of the twentieth century at a time when the geology of large areas of the State was still largely unknown. The manner of his achievement and his distinguished contribution to the advance of geological knowledge in South Australia exemplify Mawson as the explorer-geologist. D. W. Corbett, Honorary Research Associate, South Australian Museum. North Terrace, Adelaide, South Australia 5000. Douglas Mawson was a geologist with a broad scientific background and diverse interests within his discipline, as a review of his publications will verify (see Innes and Duff 1990). Known and honoured as one of the great Antarctic explorers, the prime motivation behind his achievement was the pursuit of scientific knowledge, which he coupled with the innovative use of the latest available technology. With the exception of Otto Nordenskjold, geologist and leader of the Swedish Antarctic Expedition (1901-03), Mawson was the only science-trained expedition leader in the heroic age of South Polar exploration and his approach marked the beginning of the modern age of Antarctic exploration. Mawson's career as a geologist and his scientific reputation were built in South Australia during his many years in the geology department at the University of Adelaide. While this paper focusses on the academic aspects of Mawson's life more than his Antarctic exploits, it is impossible to disentangle the two strands of endeavour. Down south he was the scientific- explorer; in the arid outback of South Australia he was the explorer-scientist. His major researches inextricably linked the two vastly different environments, for it was the opportunity to study an ice-age at first-hand and to equip himself better as an interpreter of past glaciations that inspired Mawson to venture south with Shackleton in 1907, The circumstances of time and place (early twentieth century South Australia, much of it little known geologically) and his life-long interests in glaciation, ancient and modern, determined that Mawson the geologist and Mawson the explorer were one and the same and through his extensive investigations in South Australia he personifies both the adventurous and the enquiring spirit. Mawson's Antarctic exploits have been chronicled by historians, fellow-expedition members and in his own classic account 'The Home of the Blizzard'. His career as a scientist, and notably an assessment of his geological work, have been less well treated, although aspects of them are covered in the biography by his wife (Mawson 1962) and by Alderman (1967); by Parer and Parer-Cook (1983); in the biographical summary by Jacka (1986) and by Sprigg (1986). This paper presents a review, with emphasis on Mawson's career as a geologist in South Australia and his contribution to geological knowledge of the State. Early Years As one of the first generation of geologists trained in Australia, Mawson brought an Australian rather than European perspective to his work. Although born in Yorkshire, England on the 5th May, 1882, he was only two years old when his family migrated to Sydney and he grew up staunchly independent and thoroughly Australian in outlook— as the promotion of his Antarctic expeditions was to demonstrate. He 108 D. W. CORBETT 14° 16° 18 ( Torres Islands \ \ Q»° Banks Islands O Espirit' Santo tu l ^A ™. IWaewo SYDNEY Pentecost I. Malekula CT^S Ambrim o£> Efate 100km % \^t\ \. j Eromanga Tana -< CJ c -E « - • *- 1 ** \. e w , *«- •— K« i^d K£ ZC o-- \\ 3 Z D c m £** fc- O S-J t> o CO o III _ Z SO « z 05 I- ■ o Of I- z < CO fe fa] I X D O (/) in o CO _ '3 o s c c c o 112 D. W. CORBETT Although disappointed at not reaching their objective, Mawson collected specimens and made sketch sections of the rugged hinterland (Fig. 3). Fieldwork proved more productive along the coast where three separate pyroclastic units intruded by a variety of igneous rocks were found. An earlier report of older basement rocks (gneiss) in this part of the island could not be confirmed despite diligent searching. The northern-most islands in the chain were covered in a cursory reconnaissance, but enough to confirm the volcanic nature and essentially basic composition of the Banks Group. Arriving back in Vila in mid-August after four months away, the expedition was greeted by a minor earthquake with another tremor following a few days later — enough to indicate the seismic instabilty of the region. Finally, investigations in the Havannah Harbour area provided key sections in the local sequence of volcanic sediments and overlying reef deposits. The New Hebrides expedition was inevitably very much a reconnaisance and Mawson was keenly aware of the limitations of his survey. One of the strengths of his published report (Mawson 1905) was his inclusion of all previous references, both published and unpublished, to the geology and physical features of the islands. The general seismicity of the region, already well documented, and which Mawson himself experienced, he attributed to local volcanic sources, recognising also the importance of regional tectonic movements along major faultlines. He wrote: 'before long numerous seismological stations will be distributed through the South Pacific Islands, as only by analysing such data as would result, can definite decisions regarding the present earth- movements in this much-troubled area be arrived at*. It was to be fifty years before that hope began to be realised. An anticipation that because the old rocks of nearby New Caledonia were mineral-rich, the New Hebrides might be similarly endowed was quickly dispelled because of the geological youthfulness of the islands. Sulphur deposits and iron-bearing sands were the only resources considered to be of possible future commercial value. Within the regional context of the Pacific basin, Mawson viewed the New Hebrides island arc as of late Tertiary age, one of a series of parallel arcs which were remnants of a much fragmented continental mass. Folding and over-thrusting, associated with volcanicity, against a rigid foreland lying to the west (New Caledonia), had been followed by major faulting and sea- floor subsidence. The evidence for elevation which most impressed Mawson were the raised coral reefs. These capped the Tertiary deposits and ranged in terraces rising to over 2 000 feet (609 metres). They were estimated to be post-Pliocene in age, and gave a clear indication of continued tectonic activity since the middle Tertiary. His published account of the New Hebrides expedition put the islands on the map geologically and established Mawson as a resourceful and perceptive field worker, competent in many areas of geology and with a flair for synthesis. He foreshadowed many of the more recent interpretations of south-west Pacific region geology made in the light of Plate Tectonic theory, with the New Hebrides viewed as an island arc of volcanic lavas and pyroclastic rocks and intrusions adjacent to a deep ocean trench and above a steeply dipping seismic zone. Early Researches On returning to the University, Mawson was soon involved in research (with fellow student and later Antarctic explorer Griffith Taylor) on the igneous rocks in the Bowral-Mittagong area, south-west of Sydney. The Mittagong study (Taylor and Mawson 1903) was the first of many on the petrography and petrology of igneous rocks in their regional setting which was to remain a prime interest in the years ahead. About this time Mawson also began investigations into the radioactive properties of minerals. This was a pioneer study in Australia, closely following the discovery of radioactivity by Becquerel in 1896 and the follow-up work by the Curies which led to the isolation of radium in 1898 and the recognition that uranium and thorium-bearing minerals were radioactive. Mawson reviewed these recent researches and carried out tests on natural waters, mud, soil and air. He was required to construct his own analytical equipment, notably a gold-leaf electroscope, and a large number of Australian minerals was tested including the only two known occurrences of uranium-bearing minerals in Australia and which gave the predicted high readings. The other minerals found to test positive were mainly monazites from various tin- fields around Australia. His short paper (Mawson & Laby 1904) marked the beginning of a life-long interest in the geological occurrence of radioactive minerals and their economic potential. DOUGLAS MAWSON GEOLOGIST 113 South Australia 1905 Late in 1904, Mawson, with the support and encouragement of David, was appointed to the position of Lecturer in Mineralogy and Petrology at the University of Adelaide. He arrived to take up his academic duties full of enthusiasm and with high expectations but was immediately confronted with a lack of suitable accommodation and facilities for geology due to the University's chronic shortage of funds. Frustrations developed which were to dog him for many years, including the necessity to forge a working relationship with the only other member of the Geology Department, Walter Howchin. Already with an established international reputation and with a great deal of experience of South Australian geology, Howchin was 60 years of age when Mawson, 23, arrived on the scene. It seems there were tensions from the start. The question of 'spheres of activity' (Mawson's phrase) provided an immediate and ongoing irritant, to be temporarily removed when Mawson went south in 1907. The geology of South Australia was known in broad outline at the beginning of the century (Corbett et al 1986) This was due predominantly to the work of three men: Ralph Tate, the first geologist appointed to the University (1875- 1901); Walter Howchin, Methodist Minister and enthusiastic amateur until his appointment to the University after Tate's death in 1901; and H. Y. L. Brown, the first Government Geologist who was to act as a one-man Geological Survey from 1882 until 1912. While Tate and Howchin worked mainly, though not exclusively, close to home in the Mount Lofty Ranges, Brown had been carrying out some epic geological explorations in the northern regions of the State which at that time included the Northern Territory. He published the first geological map of South Australia in 1883, revising it periodically. His 1 899 map was a remarkable achievement and in its essentials remains little changed today. Tate had discovered glacial features at Hallett Cove in 1877; Howchin subsequently demonstrating the widespread evidence for this glaciation outside the Adelaide region and establishing a Permo-Carboniferous age for the event by about 1900. In 1901 Howchin, as mentioned previously, had discovered the much older glacials in the Sturt Gorge, a few miles south of Adelaide. But perhaps the most significant of all the early breakthroughs was the finding of Cambrian fossils in the old rocks on the western flank of the Mount Lofty Ranges (Howchin 1897). Howchin thereafter concentrated much of his effort on determining the stratigraphy of the western ranges. It was Howchin' s clearly expressed 'prerogative' to work in most areas of the State that prompted Mawson to choose the remote Olary-Broken Hill area for his first extensive field investigations. It was a decision both judicious and timely. Not only did this work extend Howchin' s researches on the supposed Cambrian rocks of the Mount Lofty Ranges into the Olary Spur and over the border to Broken Hill and the Barrier Ranges, but the problems presented by the region also satisfied his interests in mineralogy, petrology, and stratigraphy and his discoveries were to influence his future involvement with glacial deposits both ancient and modem. The explorer-geologist was severely tested in the harsh environment near the north-east state border. It was (and is) an arid, inhospitable and sparsely inhabited country. Previous investigations had centred on and around the Broken Hill ore- body itself — the largest silver-lead ore body in the world. While little was known of the origin of the ore-body and its stratigraphic controls, Mawson set himself the task of ascertaining the broader regional context of the deposit. His methods involved covering the ground in a series of traverses on foot, horseback, bicycle and on occasions, by motor vehicle. He was to determine and define two major rock divisions in the Olary- Broken Hill region: 1. An older basement complex of metamorphic rocks including gneiss, schist, amphibolite and granite of Precambrian age which he named the Willy ama Complex, and 2. A younger series of sedimentary rocks, in part metamorphosed, of assumed Cambrian age including a wide variety of rock types, the most notable and persistent being boulder-beds and related sediments of glacial origin. This he named the Torrowangee Series. In January 1907, the Australasian Association for the Advancement of Science (AAAS), held its meeting in Adelaide and Mawson gave two (unpublished) papers based on his recent field work and particularly concerned with the glacial beds. His interest in these beds and a desire to reach a full understanding of their nature and origin was developing rapidly. Many years later he wrote to Edgeworth David that his commitment to glacial geology dated from December 1906 when his Olary field investigations were well underway. Matters moved very rapidly at the end of 1907. Mawson heard, through David (who was 114 D. W. CORBETT to join the British Antarctic Expedition), that the leader, Ernest Shackleton, was to visit Adelaide briefly on his way to New Zealand to join the 'Nimrod'. Mawson successfully requested a meeting and volunteered his services for the summer season. Within three days Shackleton had offered Mawson the position as physicist for the duration of the expedition. The opportunity to realise his ambition to experience 'an ice-age in being' was too good to miss. He accepted, quickly arranging for a ' locum' to cover his absence from the University, and left Adelaide before the end of December bound for the Antarctic. South With Shackleton - the British Antarctic Expedition David was senior geologist on the expedition with Raymond Priestley, a young man with no formal qualifications in science, as his deputy. Mawson as physicist had no official role in this area. Yet from the start David involved his old student in the geological program and in two of the three main journeys undertaken by the expedition — the ascent of Mount Erebus and the trek to the South Magnetic Pole. Mawson played a crucial role, both as a young and vigorous team member and a valuable contributor to the geological work, which David was quick to acknowledge in his final report (David & Priestley 1914). Here Mawson is referred to as Mineralogist, Chemist and Physicist to the expedition (the study of Antarctic ice — its crystallisation, granulation and other properties- — was considered by David as much a matter of mineralogy as of physics). The report incorporates much of Mawson's thinking and his contribution to it was wide-ranging, covering mineralogy and petrology and stratigraphy, as well as economic geology implications and broad- scale considerations of Antarctic geology within a global context. Mawson and David were both in the small party which reached the summit of Mount Erebus (3 794 m) on 8 March. Mawson made a traverse map of the eight day journey which provided its share of adventures (Shackleton 1911). He also took photographs and collected rock specimens. Later that winter (June 14) the volcano was seen in eruption from their base, steam issuing from the crater, although no lava was observed on the flanks of the cone. With the coming of spring, preparations were made for the two major assaults of the expedition; Shackleton's push for the geographic pole and an attempt by a northern party (David, Mawson and McKay) to reach the South Magnetic Pole. The route was along the coast of Victoria Land as far as the Drygalski Glacier Tongue, where they were to turn inland climbing onto the Polar Plateau towards the pole. A general survey of the geology of the coast and where possible of the Western Mountains was an important part of the program. The early stage of the journey was along the edge of the sea-ice with outcrops confined to occasional rocky bluffs projecting from under the ice-cap. Observations were made at Cape Bernacchi (marble), Cape Irizar, Granite Harbour and Depot Island. The rocks encountered were highly altered (metamorphic) and included gneiss and schist together with granite (itself metamorphosed) and basic lavas. Looking inland through field-glasses, these lavas were seen at a higher level to overlie the granites and in turn to be succeeded by stratified rocks which were taken to be the Beacon Sandstone, discovered and described by Ferrar (1907) on Scott's first expedition (1901 03). David described Depot Island as \.. most wonderful geologically and a perfect elysium for the mineralogist'. Mawson relished the variety of rocks and minerals they found, among them epidote, pyrite, copper pyrites (chalcopyritc), garnet and manganese oxides, some of which he thought might prove of economic value. Certain of the rocks, notably the gneiss and granite reminded him of those in the Broken Hill area. Meanwhile the ubiquitous glacial debris (moraine) dumped at the melting ice-front was a rich source of supplementary evidence of what lay further inland, below the ice-cap. Erratics at one locality were very similar to the Victor Harbor Granite. The ascent of the plateau became an ordeal and any thoughts of geology were replaced by concentrating on the dash for the magnetic pole. The elusive and ever-wandering point was reached on 16 January 1909. The return journey proved a battle for survival and the party was very fortunate to make the rendezvous with the 'Nimrod' and safety. The information collected by the Magnetic Pole parly made a valuable contribution to the geological results of the expedition. Notably, David and Mawson were able to build on Ferrar's earlier work and show that Victoria Land comprised a basement complex of metamorphic, igneous and mineral-rich rocks of assumed Precambrian age, overlain by a sedimentary sequence, which, on the limited evidence available, included rocks intermediate in age DOUGLAS MAWSON GEOLOGIST 115 FIGURE 4. Mawson, seated, on his return from Antarctica, 1909. Howchin stands at his right shoulder. Photo: P. Driver-Smith. between the basement and the Beacon Sandstone. For Mawson the expedition proved a stimulating and scientifically rewarding experience. His desire to experience modern glacial processes in action had been achieved and the insights gained were to be developed on his next journey south and later applied to the ancient glacial deposits of South Australia. Above all, his participation in the Shackleton expedition had determined his lifelong interest in the southern continent. Interlude Mawson returned to Adelaide in April 1909 to a hero's welcome (Fig. 4), but within a few weeks he was back in the Olary-Broken Hill country, hiring a horse and dray and continuing his investigations among the mineralised rocks and the ancient glacial beds of the arid north. After completing his fieldwork, he spent some time in Broken Hill writing up his results, and on laboratory studies of the mineralogy and petrography of the key rock types. He was particularly impressed by the pegmatites which were such a prominent feature of the region and wrote: 'In no other part of the world can pegmatite formations occur on a more extensive scale'. The study of pegmatites was at that time in its infancy and Mawson recognised that in Olary he had evidence of their importance as 4 a connecting link between the igneous rocks and the differentiated products economically worked as ore deposits'. His fieldwork, and the subsequent published review (Mawson 1912a; Fig. 5) of all aspects of the geology and economic geology of this important and yet little known mineral province, was a major contribution to South Australian geology and earned him the degree of D.Sc. from the University of Adelaide in 1909. While in Broken Hill he examined and valued several mineral collections from the mines on behalf of the South Australian Museum. Some of them included specimens of outstanding quality. Mawson had first acted for the Museum in 1906, when he had successfully acquired the Dunstan Collection from the Wallaroo Mines. In 1908 he was appointed Honorary Curator of Minerals, thus beginning a long and productive association with the Museum which was to last his lifetime, and which became crucial in later years, as no full- time curator responsible for the collection was appointed until 1956. The Mawson-Museum connection cannot be treated in any more detail here. Early in December 1909, Mawson left Adelaide for London on board the SS 'Mongolia'. He was 116 D. W. CORBETT FIGURE 5. MawsoTTs interpretation of the geology near Olary, South Australia (1912b). Courtesy ANZAAS. keen to go to the Antarctic again and full of ideas for a scientific program that he hoped would appeal to either Shackleton or Scott, both of whom he knew had plans for expeditions south. Scott was very keen to enlist Mawson (even offering him a place in the final polar party), but was not prepared to consider any modification of his own plans which were already well advanced. Discussions were long and frank but when Mawson learned that the position of chief scientist (which he might have accepted) had been filled and his proposal for a western party to explore King George V Land unacceptable, he declined Scott's offer. He was simply not prepared to compromise his purely scientific approach to Antarctic exploration. As for Shackleton, initially Mawson maintained friendly relations with his former leader, even visiting a gold prospect in Hungary which Shackleton was convinced would make his fortune (and Mawson's too if he came in with him as an investor). But becoming increasingly disillusioned with Shackleton's erratic behaviour and lack of firm plans, Mawson followed him to America in an attempt to pin him down; and the two signed an agreeement that an expedition would leave late in 1911 and should Shackleton be unavailable as leader Mawson was to take over and receive all Shackleton's support and assistance in raising funds. Ever the realist, it was clear to Mawson that the responsibility for mounting the expedition was most likely to be his and so it proved. He was then faced with the daunting task of raising funds quickly and in competition with Scott who had a head start and the advantage of being based in London. Mawson returned to Adelaide. For the next eight years the Antarctic was to dominate his life. But for a brief interlude he returned to academic life and his research activities and to these we now return. Radioactive Minerals: Radium Hill and Mount Painter Mawson's involvement with radioactive minerals had been renewed soon after his arrival in South Australia when in May 1906 carnotite, a highly radioactive mineral, was found in ore taken from a mine 24 miles east-south-east of Olary. H. Y. L. Brown visited the site and issued a report in which he speculated that the mineral was a decomposition product of primary uranium ores occurring at depth. In August, Mawson obtained samples which he passed over to E. H. Rennic, Professor of Chemistry, for analyses which DOUGLAS MAWSON GEOLOGIST 117 s g FIGURE 6. Geological cross-section in the Mt Painter Arkaroola area (1912b). Courtesy ANZAAS. suggested that black minerals initially identified as magnetite, were far more complex. This prompted Mawson to visit the mine and make a detailed survey of the exposed reefs which had been pegged originally in the belief that the prominent black mineral was a tin-bearing ore. Further examination showed that while the magnetic iron/titanium mineral was dominant in the main reef, a rarer black mineral with a distinctive and very brilliant lustre also occurred as grains, streaks and cubic crystals. Analysis shows that in addition to a high percentage of iron and titanium the mineral also contained rare earth elements together with uranium, vanadium and chromium. Mawson proposed the name Davidite for what he believed to be a new mineral species in honour of his former teacher (Mawson 1906). He also identified Roscoelite, a bright green vanadium-bearing mineral. At a time when knowledge of radioactive minerals was in its infancy, Mawson recognised the deposit as a significant discovery, possibly of economic importance, and named the locality Radium Hill. He was optimistic but cautious in his appraisal: 'This body of radio-active ore is, in the matter of quantity, much the most important yet discovered in Australia. Its low grade, however, introduces serious difficulties to commercial enterprise in this direction'. Radium Hill was to have a long but intermittent history based on the working of its radioactive lodes (Fig. 7), and Mawson was to renew his interest in the deposit in the 1920s. Mawson visited the Flinders Ranges for the first time towards the end of 1910, during his short spell in Adelaide between Antarctic expeditions. He had received a visit from W. B. Greenwood, the owner of Umberatana Station, who had brought some specimens of a bright green mineral which Mawson identified as torbernite, an hydrated phosphate of copper and uranium. Greenwood had been prospecting in the northern ranges on behalf of the Mines Department for several years and had become annoyed when specimens, including the green mineral, which he had left at the Department some time before had not received attention; whereupon he had retrieved them and taken them to the University. Excited by the prospect of another radioactive mineral deposit in South Australia, Mawson headed north and, with Greenwood and renowned prospector Harry Fabian, made an examination of the geology and mineral occurrences in this most rugged and inaccessible part of the Flinders, attempting to place the local geology in the broader regional context (Fig. 6). The short visit marked the 118 D. W. CORBETT beginning of his long and fruitful association with the geology and mineral resources of the region and his field notebook for the trip is full of detail of his observations and the excitement of discovery; for example: 'The results of my recent journey have been in the highest degree satisfactory... My main objective was the investigation of the metamorphic rocks eastward of Yudnamutana in connection with the interpretation of the Barrier Ranges, a subject in which I have long been absorbed. . . ' With regard to the potential of the uranium- bearing lodes he wrote : '... the outcrop is on the whole low grade, though richer patches are met with at intervals. Improvement may be expected below. The ore can be treated very inexpensively and this will offset the low-grade character. So far as I am aware this is the most extensive uraniferous lode formation in the world'. The back of his field notebook contains estimates of costing for ore extraction, transport, freight and handling, including comparisons for treatment on the spot, at Port Pirie or in England. There would be a need for roads and wells and for camels to bring out the ore. If the grade proved high enough, he foresaw the setting up of a company to work the ore — the Radium Extraction Company. He even worked out the capital required - £20 000. These estimates may have been doodlings around the campfire, at the latest they were put on paper soon after his visit, and they have an immediacy which reveal Mawson's keen interest in the region, its economic potential and his desire to be part of the action. He was not to return north for more than a decade. Over the next few years some intermittent mining activity took place at Mount Painter, but with the outbreak of World War I activity ceased. South Again - The Australasian Antarctic Expedition (AAE) 1911-14 In January, 191 1, Mawson gave a lecture to the combined geology and engineering groups at the annual meeting of the AAAS in Sydney. It was a fund raising exercise designed to appeal to national pride as well as pointing out the scientific and future economic potential of the southern continent. He stressed that: 'the collection of scientific data (was) obligatory upon us ... its mountains were blessed with mineral resources ... (and) it holds among its fossil strata secrets especially interesting to Australians'. Science was to be linked to technology and the latest developments — the wireless and motor sledges — still to be tested down south, were seen as a vital aid to scientific exploration. The burdens of leadership were enormous leaving little time for any geological researches of his own, but Mawson recruited a strong team of four geologists who were to carry out valuable surveys over a great expanse of territory. Cecil Madigan, later to join his leader at the University of Adelaide, was appointed meteorologist. Reaching the area selected for the expedition's base, the expedition ship 'Aurora' had great difficulty in breaking through the ice barrier and was forced to cruise westward; Mawson had hoped to find the Antarctic continent in these latitudes bounded by a rocky and attractive coast like that in the vicinity of Cape Adare, the nearest well-explored region. The poor prospects of carrying out extensive geological work under such conditions must have been a blow, but Mawson was impressed and awed by the sheer physical presence and power of the ice and its dominance of the environment: 'The land was so overwhelmed with ice that, even at sea-level, the rock was all but entirely hidden. Here was an ice- age in all earnestness; a picture of northern Europe during the Great Ice Age some fifty thousand years ago' (Mawson 1915, 1996 p. 40). The main base was established at Cape Denison (Commonwealth Bay) and a western base party, including two geologists, some 1 500 miles (2 400 km) further along the coast to the west on the edge of the Shackleton Ice-Shelf. The vicinity of Cape Denison was comparatively ice-free 7 and the geologists were able to study glacial erosion and deposition at first hand. The rocks outcropping near the main base were predominantly metamorphic — folded and contorted gneiss and schist of presumed Precambrian age. Evidence of ore-bearing minerals was also found, among them iron, copper and molybdenum which Mawson believed gave promise of larger bodies in the vicinity and indicated the probability of mineral wealth beneath the continental ice-cap. Around the edge of the ice margin, the terminal moraines were 'a veritable museum'. Rocks, showing every variety in colour and form, were assembled, transported from far and wide over the great expanse of the continent ... The story of the buried land to the south is in large measure revealed in the samples brought by the ice and so conveniently dumped' (Mawson 1915, 1996 pp. 74-76). After a particularly severe winter with hurricane force winds, the summer journeys commenced, and much valuable scientific work carried out. A western party made a particularly DOUGLAS MAWSON GEOLOGIST 119 interesting find, the first meteorite from Antarctica, duly named the 'Adelie Land Meteorite 7 , and the first of many to be found in later years. Mawson himself, with Ninniss and Mertz and the dogs, set off on the longest and most ambitious of all the journeys, striking east on a southerly course in an attempt to cover as much territory as possible in the time available. The story of that harrowing journey which cost the lives of both of Mawson's companions, and of his miraculous return against overwhelming odds just in time to see the 'Aurora' disappear over the horizon, is an epic story of endurance, fortitude and sheer will power which has been well told by Mawson himself in L The Home of the Blizzard' (Mawson 1915) and by Bickel (1977). Mawson and the small remnant of the expedition left behind to search for him had to face another Antarctic winter. The 'Aurora' returned in December 1913 and on the 26 February 1914, Mawson arrived back in Adelaide. The War Years and After The expedition was home, but for Mawson there was to be no respite and it was to be a further six eventful years before he was able to resume an ordered academic life again. The First World War began as he was returning from an overseas business trip to Europe concerned with the wind-down of the expedition, and by 1916, after travelling to the United States, he was back in Britain where he spent the remainder of the war on highly responsible government work with the Ministry for Munitions. He did not return to Adelaide until May 1919. Back at the University he found that little had changed in his absence except that Walter Howchin had been awarded the title of Honorary Professor in Geology and Palaeontology in 1918. Mawson (1944a) wrote later: '1 was not a little confounded. 1 would have packed up and left immediately but for the supreme consideration of faithfully recording and adequately publishing the immense mass of data received during my Antarctic Expedition'. Early in 1920 he talked with Howchin about the future: 'stating that I was not prepared to continue at Adelaide for another year under the then existing inadequate arrangements for geology'. There was no prospect of the much-needed new building for geology and Mawson tied himself to a compromise twenty year agreement on additional accommodation which proved increasingly inadequate as years went by, and which he was later to regret. Meanwhile, Howchin's continuing presence at the University was perceived by Mawson as a hindrance to the satisfactory development of the Geology Department and it must have been a great relief to him when Howchin, then 75, decided to retire at the end of the year. In 1 92 1 Douglas Mawson was appointed to the newly created Chair of Geology and Mineralogy, and Dr Cecil Madigan, his old comrade from the Antarctic, joined him as Lecturer the following year. Mawson and Madigan were to be the only two staff members in the Geology department for much of the inter-war period. The establishment of the new department and his teaching and administrative duties (not to mention continuing Antarctic matters) took precedence in the early 1920s but field investigations were resumed in both the Olary region and the north Flinders Ranges. While he built on his work in the Boolcomatta area, locating several key sections spanning the junction between the basement and overlying sedimentary sequence, he was never to resume a sustained fie Id work program in the region, despite the urgings of David. David's retirement at the end of 1923 and the politics involved in choosing his successor proved unsettling for Mawson who at that time was still not fully committed to staying in Adelaide and applied for the job, and although the best qualified of the candidates, he was ultimately unsuccessful because of David's influence on the course of events and his (perhaps over-zealous) loyalty and support to Leo Cotton, his deputy and locum for many years, who was appointed to the Chair. The outcome certainly posed strains on the David- Mawson relationship, and further tensions and distractions were to follow when David's increasing obsession with the search for supposed Precambrian fossils focussed on the old rocks in South Australia (see also pl24, End of an Era). On an earlier visit to Adelaide in 1921, David had found what he believed to be arthropod remains in the old rocks at Reynella (long considered by Howchin as Early Cambrian) tentatively placing them in the late Precambrian (Proterozoic) and proposing a new division the Adelaide Series. Meanwhile, Mawson turned his attention to the north Flinders once again. He had published a brief comparison of the Mount Painter and Olary provinces (Mawson 1912b), but the results of his investigations around Mount Painter in 1910 were treated more fully in a long-delayed paper (Mawson 1923) in which he recognised a core of 120 D. W. CORBETT :•' MM FIGURE 7. The Radium Hill Mine, soon after its discovery. Mawson, 1944b, pi. 19. Courtesy Royal Society of South Australia.. * #:i FIGURE 8. Mount Gee, viewed from the west. Mawson, 1944b, pi. 19. Courtesy Royal Society of South Australia. DOUGLAS MAWSON GEOLOGIST 121 Precambrian igneous and metamorphic rocks overlain by a sedimentary series similar in many respects to those overlying the basement in the Adelaide region. He followed David's lead in applying the term Adelaidean (Adelaide Series) to these sediments and suggesting a Proterozoic (late Precambrian) rather than Cambrian age. Igneous activity included two acid phases, the second with pegmatites and aplites followed by a volcanic phase characterised by vesicular basalts. The mineralisation of the Mount Painter-Mount Gee area, notably the quartz reef at Mount Gee and the uranium minerals (torbernite and autunite) at Mount Painter, was described in detail (Fig. 8). Following Mawson's enthusiastic report on the uranium potential of the area after his initial visit in 1910, several companies had been floated and prospecting carried out on Radium Ridge and Mount Gee. Meanwhile the Mines Department had been more cautious but in favour of further prospecting. Sporadic attempts at developing the prospects (notably the No. 6 workings) had languished during the war, but in 1924 there was renewed interest in radioactive ores world-wide and a return to the Mount Painter field seemed justified. Consequently, Mawson returned north in November 1924 and spent much of his time in the Mount Painter area on detailed study of the mineralised rocks but also in broader scale assessment of the regional context of the mineralisation. The regional geology of the Flinders Ranges was little known at this time. Howchin had made a pioneer traverse from west to east at the level of the Parachilna Gorge in 1907 but the results were not published until 1922. While on his return trip from the north, Mawson had the good fortune to make a particularly significant discovery just west of the Italowie Gorge. It was a case of the prepared mind recognising the significant evidence and provides a good example of Mawson's versatility as a geologist, for fresh from the mineral riches of the Mount Painter country he was now to show that he had a good eye for a FIGURE 9. Entenng Yudanamutana. Mawson, 1923, pi. 33. Courtesy Royal Society of South Australia. 122 D. W CORBETT fossil also. What he found were 'curious markings' on roadside outcrops of limestone which 'on inspection ... was found to be due to massed fossil heads of a Cryptozoon-Wkt alga ... developed in a massive formation, the algal remains constituting the bulk of at least some of the beds' (Mawson 1925). Similar structures had been reported from the MacDonnell Ranges by Chewings and described by Howchin (1914) but there was as yet no general agreement on their stratigraphic position. Mawson inclined towards an Early Cambrian age, but David, on hearing of the find thought they might be older (Proterozoic). Very little was then known of Precambrian life forms; most of the obscure remains described being referred to as Crpytozoon (hidden life)— but the recently announced discovery of algal fossils in Precambrian rocks in Montana by Walcott suggested that the Flinders rocks might be of similar age. David certainly believed so and urged Mawson to investigate further. Mawson's 'Cryptozoons' are now recognised as algal stromatolites, These columnar and dome-shaped structures produced by the trapping of sedimentary particles by mats of blue-green algae are the earliest abundant fossils and are well developed in certain of the Precambrian rock formations in the Flinders Ranges. Mawson, however, was not to be pressured into narrow specialisation and continued his wide-ranging investigations of Flinders Ranges geology (Figs 9, 10), publishing further papers in the mid 1920s. One was on the volcanic rocks of the Wool tana area (Mawson 1926a), first encountered in Arkaroola Creek in 1910 and subsequently examined near Wooltana on a brief visit in 1924. There he found a thick sequence of volcanics — vesicular lavas, tuffs and agglomerates overlain by boulder beds and finer sediments which he described as: 'most probably of fluvio-glacial and glacial origin'. Large granite boulders higher in the section were even more convincing, leaving: \.. no doubt in my mind ... as to the glacial nature FIGURE 10. The quartzitc of the Yudanamutana Gorge. Mawson, 1923, pi. 33. Courtesy Royal Society of South Australia. DOUGLAS MAWSON GEOLOGIST 123 of these beds,..' Sediments above the glacials, unsuccessfully searched for fossils, were believed of likely Late Precambrian age. Another spin-off from the 1924 North Flinders trip was a visit to the Paralana Hot Springs which interested Mawson because of the radioactive nature of the waters (Mawson 1927a). For a short period about this time an enterprising doctor had sought to capitalise on the supposed therapeutic qualities of the spring by setting up a spa on the site, although with no success. The springs are situated close to the faulted junction between the old rocks of the ranges and the younger sedimentary basin to the east. While most geologists supported an origin from deep-seated basin waters, Mawson speculated on a possible contribution from the highly uranium-rich rocks lying to the west. He collected water and gas for examination, hoping the latter would prove rich in helium. Unfortunately the samples did not survive the journey back to Adelaide. The highly active gas which at first puzzled the analysts, has since proved to be rich in radon; the water is much less radioactive. Mawson had made a brief visit to the Willouran Ranges on the north western flank of the Flinders in 1920. From here he described and named the Willouran Series—a thick sequence of sandstones, slates and calcareous beds underlying a dominant quartzite formation and forming a synclinal structure (Mawson 1927b). Sediments of probable fluvio-glacial origin were also found in the eastern part of the area. Mawson's activities in the Flinders Ranges during the 1920s were essentially of a reconnaissance nature and based on relatively little time, overall, spent in the field. The information included in the papers cited above demonstrate how much he could gain from a brief investigation, displaying his strength as a field worker and skill in rapidly appraising the geological situation in unknown country, and revealing his pre-eminence as an explorer- geologist. Work carried out during this period, while not yet part of a coordinated research programme, provided the basis for his later systematic approach to regional studies in the 1930s and 1940s. Away from his empirical work in the field, he demonstrated a comparable ability to synthesise existing knowledge in a lengthy treatment of the igneous rocks of South Australia, delivered as a Presidential Address at the AAAS meeting at Perth in August 1926 (Mawson 1926b). This was an important review, given the widespread distribution and importance of igneous rocks, notably in the older terrains of the State, but also in those of younger age. By discussing them region-by-region in their structural and stratigraphic context, he produced what was essentially a geology of the State. One issue of particular interest, not to be resolved for many years, was the age of the highly metamorphosed sediments on the eastern flank of the Mount Lofty Ranges and their associated intrusives. With reference to Howchin's view that these rocks were the equivalents of the Adelaide Series, Mawson urged caution, pointing out: 'the geological mapping of the Ranges is yet only in the early stages (and) final conclusions are unsafe until more detailed petrology and complete geological mapping have been accomplished'. As for the Victor Harbor granites Mawson comments perceptively: 'there is yet no convincing evidence against the assumption of an early Palaeozoic age for this granite'. Central Australia 1927 In November 1927, Mawson and Madigan made a short visit to Alice Springs and the Western McDonnell Ranges to check a mineral occurrence which had excited local interest. A white crystalline substance, long known to the aborigines and which burned with a fierce flame, was thought possibly to have some connection with petroleum. Specimens had been brought to Adelaide (initially to Madigan), and identified by Mawson as potassium nitrate (saltpetre or nitre) a rare mineral with some economic use in the chemical and explosives industries. Realising the opportunity for geological exploration in an area still little known geologically, Mawson quickly organised an expedition to investigate. They travelled north by train to Alice Springs, and their car journey out through the western ranges from Alice Springs blazed a trail through country never traversed by vehicle before. The deposit proved to be of organic origin, the long-time accumulation of animal droppings, of scientific but not economic interest. Much of the route west ran along the strike of the well-exposed sedimentary beds which gave a general indication of the southward dipping succession between the Precambrian crystalline basement and known Ordovician rocks of the Horn Valley. Traverses were run and rough sections made at several localities along the way, including Ellery Creek, and over 10 000 feet of strata were measured. In the lower beds 124 D W. CORBETT 'cryptozoal limestones' and organic structures similar to those from the Flinders Ranges were prominent in several areas, while worm tracks and traces of mollusc shells together with algal limestones distinct from those lower in the sequence were found in the younger beds. On the basis of his limited field evidence, the sedimentary sequence was divided into two divisions — the Pataknurra and Pataoorta Series, once again showing his flair for the rapid assessment of the general geological situation in a new area. He made tentative and essentially correct correlations between the central Australian rocks and the older sequences in South Australia. This first and, as it proved, only venture into central Australian geology was a major contribution to knowledge of the region, building on and revising the earlier work of Tate and Watt (on the Horn Expedition), H. Y. L. Brown, C. Chewings and L. K. Ward. Mawson visited London the following year where he exhibited photographs and specimens at the Geological Society. The results of the expedition were published by the Society (Mawson & Madigan 1930). While circumstances determined that Mawson never returned to the Centre, it seems likely that he always envisaged following up the promising results of his first trip. Madigan, meanwhile, had been captivated by the Centre and immediately began his own independent and wide-ranging researches into the geology of the region which were to produce some of his best work. This perceived 'take-over' by Madigan was resented by Mawson and led to bad feeling between the two men, but Mawson was soon to be occupied with other matters. BANZARE 1929-31 Mawson was very much involved with Antarctic matters from 1929 through into the early thirties. BANZARE (The British, Australian and New Zealand Antarctic Research Expedition) differed from Mawson's previous expeditions in being ship-based. The program was carried out on two summer cruises (1929-30, 1930-31) and while no land bases were established, limited sorties were made ashore, notably on some of the sub-Antarctic islands. With the emphasis on marine science, much of the data was collected by systematically sounding the depth of the offshore waters, sampling the water column and dredging the bottom for biological specimens. Because of the nature of the expedition, geological work was restricted, although sediments collected from the sea-floor adjacent to the land and the variety of ice-transported boulders dropped there provided abundant evidence of sedimentation and sedimentary processes in a modern glacial environment. The BANZARE expedition further inspired Mawson to continue his study of the ancient glacial deposits in the Flinders Ranges, and he was to do this with renewed vigour in the years ahead. The extended geographical and geological knowledge of Antarctica resulting from the BANZARE voyages enabled Mawson to be the first to postulate the existence of an extensive landmass lying beneath the southern ice-cap and bounded by a wide continental shelf. Much of the evidence for the geological make-up of this buried continent came from the ocean-floor studies where dredged boulders proved to be rocks of continental type. Mawson envisaged that the mineral resources of Antarctica, the existence of which his expeditions had suggested but were yet to be proven, would be developed at some time in the future, and that underground mining operations could be carried out successfully despite the problems of climate and terrain. End of an Era Professor Edgeworth David died in 1934. In his later years he had become increasingly obsessed with his recognition of 'fossils' in the Precambrian rocks of South Australia (David 1928). Most of his finds had been made on visits to Adelaide at times when Mawson was away overseas or otherwise engaged on Antarctic business and had no chance to control these incursions. For Mawson it was particularly galling for David to hint that his old student was sitting on a rich treasure trove of Precambrian fossils while he, David, made the pickings, even enlisting Mawson's students and staff to help him collect. Wisely, Mawson held his counsel and endeavoured not to be drawn into the matter. It soon became clear, to almost all but David, that the supposed 'fossils' were in fact inorganic in origin (a verdict upheld by later workers), but he was to persist in his belief through to the end of his life, considering his discoveries in South Australia to be the highpoint of his career. Deciphering the Flinders In the mid- 1930s Mawson renewed his long- DOUGLAS MAWSON GEOLOGIST 125 UJ l ITWJ « '■ t 7 /■ $$ ' il ■!. ' *? V'r >'.-'■ < 3 o O 'u o on 13 >-. o >> ■c o U ON 1) c g to c TO i-i c 3 CQ a GO O u c o o a a. H 55 126 D. W. CORBETT u3 § *s- IS Sg &*< J I eg o o o o Q J5 00 m O DOUGLAS MAWSON GEOLOGIST 127 standing but formerly sporadic association with the Flinders Ranges. Now for the first time with fewer distractions, he was able to develop and pursue a sustained and systematic ficldwork program the results of which constitute his most notable contribution to South Australian geology. Only a brief resume of this work is given here, to be treated more fully in a later paper. The Flinders Ranges were sparsely populated in the 1930s, and moving around in the outback more arduous than today. Roads and tracks were in general poorly maintained and motor transport, then in its infancy, was still an adventurous undertaking. There were no photo-based maps or accurate large-scale topographic maps available in those days, the only maps available being pastoral plans which were quite inadequate as a base for geological mapping. Mawson's approach was to select key sections which were measured and described in detail. In a few of his papers, small- scale regional maps were included. He first used aerial photography as an adjunct to his ficldwork in the late 1940s. Field-camps in the pre- and early war years combined the instruction of senior (third year) students in field techniques, including section measuring, with the gathering of basic geological information in country previously geologically unknown. By 1939, having spent time in the Wooltana, Mount Carnarvon, Parachilna Gorge and Italowie areas, Mawson's work in the Flinders had reached a stage where review was warranted, and two publications in that year, on the Late Precambrian, and Cambrian successions respectively (Mawson 1939a, 1939b), detail rock sequences in the Brachina Creek-Oraparinna and Wirrealpa Basin areas of the central ranges (Fig. 11). The Pound Quartzite, the dominant ridge-former of the Flinders which occurs some distance stratigraphically below the oldest fossiliferous Cambrian rocks, was provisionally placed at the base of the Cambrian on lithological grounds and interpreted as marking the start of a new cycle of sedimentation. Defining the Precambrian- Cambrian boundary was to become of increasing interest to geologists in the years ahead, a task made no easier by the lack of fossils and by the absence of major breaks in sedimentation in the thick Upper Precambrian to Cambrian sedimentary sequence in South Australia. In his Wirrealpa Basin paper, two important fossiliferous units are described: the Archaeocyatha Limestone and a younger limestone (the Wirrealpa Limestone) first described by Howchin but here fully documented for the first time with its fossil fauna of algal remains {Girvanella), brachiopods (Oboiella), pteropods and trilobite fragments (Redlichia). Other papers were to follow dealing with sequences in the Aroona, Copley, Brachina and Wilpena areas and culminating in an important review, 'The structural character of the Flinders Ranges' (Mawson 1942), which brought together the results of his fieldwork program to that time. A small scale map (Fig 13) showed the location of the key sections measured and the distribution of the rocks spanning the four main periods of deposition: the Middle Proterozoic and Late Proterozoic, the Cambrian and the Tertiary- Recent. Key marker beds, including the prominent Pound and ABC Quartzites, tillites and hieroglyphic limestones were also delineated. A cross-section (Fig. 12) indicated the essentially simple anticlinal structure of the central ranges. A wider-ranging paper (Mawson 1947) extended and amplified his earlier work, adding new information from the western flank of the ranges both in the north (near Copley) and in Mundallio Creek to the south. In both areas Mawson recognised thick quartzite formations at the base of the sequence, pointing out the error made by Segnit (1939), working for the Mines Department, who had equated these ABC Quartzite sequences with the Pound Quartzite. Mawson, however, was himself misled in the area west of Quorn where he mistook the prominent folded ABC Quartzite as Pound Quartzite (here thinly developed). Such are the dangers of making correlations on lithological grounds in rocks without fossils to act as age indicators. Mawson also discussed the Flinders succession in its broader tectonic framework which he considered a major trough of deposition of Late Precambrian to Cambrian age and referred to as the 4 great synclinal basin' (later to be named the Adelaide Geosyncline). His work had demonstrated the great thickness of strata deposited in the trough which extended south from Central Australia to the Adelaide region and lay between two ancient Precambrian massifs: Yilgamia to the west and Willyamia to the cast. The flat-lying sedimentary cover-rocks to the west of Port Augusta, believed on lithological and structural grounds to be correlates of the basal quartzites in the western ranges are now known to be ABC Range Quartzite equivalents. Wider correlations of the Flinders Ranges succession with the rocks of the western Mount Lofty Ranges (comparisons with the still largely unresolved sequence on the eastern flank of the Mount Lofty 128 D. W, CORBETT THE CENTRAL AREA OF THE FLINDERS RANGES \ \ \ \ \ \ N FIGURE 13. Mawson's geological map of the central Flinders Ranges. Mawson, 3942. Courtesy Royal Society of South Australia. DOUGLAS MAWSON GEOLOGIST 129 Ranges still remain highly speculative), with Central and Western Australia and also with South Africa (Transvaal System) were also suggested and discussed (Mawson 1947). In the early 1940s Mawson returned to the North Flinders investigating the metamorphic effects of plug-like granite intrusions in the Precambrian country rocks (Mawson & Dallwitz 1945a,b). This study of both the sedimentary and igneous aspects of the geology of the region linked his early work on the older mineralised terrains in the Mount Painter area with his later concentration on unravelling the stratigraphy of the overlying sedimentary sequence, Ancient Glaciation Mawson first encountered evidence of ancient glaciation on his trip to the north Flinders in 1910. Subsequently he had demonstrated the widespread distribution of glacial rocks in the ranges during his extensive field investigations (Fig. 14). His Antarctic experience of a glaciation-in-being, with the diversity of its products and processes, from the coarse boulder beds of terminal moraines to the finely layered sediments laid down in tranquil glacial lakes, had equipped him, as he had intended it should, to recognise the varied and not always easily interpreted evidences for former ice- ages. By the late 1940s, Mawson had established the stratigraphic position of the Sturt Tillite in many areas of the Flinders and also identified a younger glacial episode in the Elatina area on the western flank of the central ranges (Mawson 1949a). The distinctive sediments— coarse sandy beds overlain by a thick unstratified boulder bed — first encountered in 1938, had initially been thought to be of pyroclastic origin (Mawson 1938) because of the preponderance of igneous pebbles in the rock. Mawson then produced evidence for a third glacial event, this time predating the Sturt Tillite (Mawson 1949b), in the deeply eroded high FIGURE 14. A large striated erratic weathered out of Proterozoic tillite, Yankarunna Station. Photo R. H. Jones. Mawson, 1945, pi. 14. Courtesy Royal Society of South Australia. 130 D. W. CORBETT country of the Bibliando Dome east of Hawker. Here a thick sequence of glacials and associated beds revealed two extended periods of glaciation separated by a well-defined inter-glacial interval. The younger glacial event was equated with the Stmt Tillite while the older (lower) tillites were referred to as the Bibliando stage. The clear demonstration of a third glacial phase in the history of the geosynclinal basin, with a broad range of deposits from boulder tillites through coarse fluvio-glacial sediments to fine laminated shales, raised such questions as the interpretation of past geographies, climates and environments, and the cyclic nature and causes of glaciation. Mawson, with his deep and extensive knowledge of ice-action both ancient and modern, was, more than any other geologist of his generation, aware of the diversity, complexity and the problems associated with interpreting glacial evidence. He discussed some of these in his paper on Bibliando, which marks the culmination of his career-long investigation of glacial phenomena. He took the opportunity to view this latest work in a broad historical perspective. The early discoveries of Selwyn, Tate and Howchin had placed South Australia at the forefront of research in the study and understanding of past glaciations. His own work had extended and clarified that knowledge, setting it within the wider context of the development of a major Upper Precambrian geosyncline, the importance of which had now become apparent. While the rocks in the Flinders are magnificently displayed and accessible for study, the lack of fossils in the ancient rocks have always made the task of the stratigrapher difficult. Mawson, the first geologist to systematically approach the task of establishing an ordered succession was acutely aware of this problem. As his work proceeded and became recognised beyond South Australia, he perceived the need for a standardisation of geological nomenclature, particularly of stratigraphic names, and he was a member of the first committee set up by ANZ AAS in 1947 which led to a Code of Stratigraphic Nomenclature being formalised in 1950. Mawson favoured dropping the long- established but restricted Adelaide Series of David and Howchin and replacing it with the Adelaide System which he sub-divided into the Torrensian, Sturtian and Marinoan Series (Mawson & Sprigg 1950). The defining of the Adelaide System marked the end of Mawson's work on the old rocks of South Australia. The discovery of jellyfish fossils by Reg Sprigg in the Pound Quartzitc at Ediacara in 1946 (Sprigg, 1947, 1949), and the later finds of a diversified animal fauna by Flounders and Mincham, led to a reappraisal of the stratigraphy of the fossil- bearing beds (Glaessner & Daily 1959). Relegation of the Pound Quartzite from the base of the Cambrian (as previously suggested by Mawson) to the uppermost formation in the Precambrian meant that the Ediacara fossils, now confirmed as the oldest animals known, became world famous. The South-East Contrasting strongly with the rugged grandeur of the Flinders Ranges, the geologically youthful and topographically unchallenging south-eastern region of South Australia attracted Mawson from the mid- 1930s. He became interested in the problematic 'coorongite', a rubbery bituminous substance first discovered in 1852 at Alf Flat east of Salt Creek at the southern end of the Coorong Lagoon (Fig. 15). A belief that the flammable material might be a petroleum product led to speculation that oil existed at depth; and several (unsuccessful) drilling attempts were made in the 1880s and again in the 1920s. By then almost all investigators had become convinced of the vegetable origin of the substance (and it has since been shown to be the product of algal blooms). Mawson, however, was unconvinced; and he inadvertently attracted overseas interest for a brief period when he took a visiting petroleum expert from Britain (Washington Gray) into the field to view the material and Gray went home to publicise the find, much to Mawson's chagrin. Some biochemical work was carried out overseas on the 'algal gel', but the matter was not pursued any further. More substantial and fruitful work in the region led to a series of publications on the small and isolated outcrops of granitic rock which protrude through the thin Quaternary cover of the Upper Southeast. Fieldwork involved exploration on a minor scale, for many of the outcrops lay in difficult country accessible only on foot or horseback and where the outcrops were not extensive. The studies were carried out in association with students or former students and covered a quadrilateral bounded by Meningie and Coonalpyn on the north and Kingston and Bordertown to the south. The field investigations and subsequent analytical work led Mawson to recognise a variety of intrusive and extrusive rock DOUGLAS MAWSON GEOLOGIST 131 FIGURE 15. Mawson, with auger, drilling for coorongite, the Coorong, mid-1 930s. Washington Gray on right. Source unknown. types in the outcropping basement: granites; adamellites and granodiorites; and quartz keratophyres (volcanic). Similarities between the intrusives of the South-East and those occurring along the eastern flank of the Mount Lofty Ranges and on Kangaroo Island supported the concept of a large granitic batholith existing beneath the sedimentary cover, and, on the available evidence. Middle to Late Cambrian in age. Finale Mawson was sixty-three at the end of the Second World War and approaching the normal age for retirement. He continued on, however, to the age of seventy and spent his last university- years rounding off his own researches and ensuring that his department was left in a healthy and secure position to meet the challenges ahead. 132 D. W. CORBETT THE MAWSON LABORATORIES SCHOOL OF GEOLOGY UNIVERSITY OF ADELAIDE WOODS BAGOT I AT BCUR.KE -SWIT H £^ I R.WI M A R. C H 1 T C. C T & . FIGURE 16. The proposed Mawson Laboratories. Courtesy Woods Bagot Pty Ltd, Adelaide. He had been acutely aware of the greatly increased role that geology would be required to play in the post-war world. In particular, well trained geologists could only be supplied by the universities which in turn would need to provide a sufficient level of support in terms of accommodation, equipment and staffing. In all these areas Mawson had been struggling against inadequacies for over twenty years. During this time, geology had made the best of sub-standard accommodation and increasingly inadequate provision for research, whereas the rest of the University had developed and prospered. It was the urgent need for a better deal for geology that prompted his memorandum to the University authorities in February (1944a). In his memorandum Mawson wrote: 'The department suffers from insufficiency of room, from inadequacy of the laboratories, and from the entire absence of any planned layout suitable for a science department. All other science departments of the University now have buildings specially planned for their needs though some are already proving too limited in accommodation for the flood of students now presenting themselves. Ours is an anachronism'. He went on to make a very strong case for the role of geology in education, its application to national development and significantly, South Australia's importance as a source of future mineral resources. He envisaged Adelaide University as a future centre for teaching and research in economic and mining geology and itemised the basic needs of a modern School of Geological Studies, estimating the cost of the new development at around £25 000. His argument was decisive. The University responded appropriately, if belatedly, and the Mawson Laboratories were constructed, largely to his specifications, between 1949 and 1952 (Fig. 16). He just had time to move into the new building before retirement. By then his staff had increased to five and included E. A. Rudd who was appointed to the new Chair of Economic Geology in 1949, and Martin Glaessner, a palaeontologist with an international reputation who was to become the authority on the Precambrian fossils from Ediacara. Arthur Alderman, former student and junior lecturer in the department, returned after a period away from the University to take over from Mawson as Professor of Geology and Mineralogy in 1953. In the years following his retirement Mawson remained active and influential in Antarctic matters and continued his long association with the South Australian Museum as Honorary DOUGLAS MAWSON GEOLOGIST 133 Curator of Minerals and Member of the Museum Board. He was appointed Chairman of the Board in 1952. Douglas Mawson's broad interests in all aspects of geological science and his unfailing encouragement and support of young people with an enquiring mind and an interest in science is well illustrated by the background to his last published paper (Mawson 1958). It deals with australites (small glassy objects of meteoritic origin) collected by a young schoolboy, Mervyn Pens, from the Murray Plains west of Morgan in the mid-1 930s. Mawson had visited the site and established a correspondence with young Pens, which led to further finds and ultimately to the public aquisition of a large and significant collection of australites now housed in the South Australian Museum. The paper was read before the Royal Society of South Australia on 10 October 1957. Sir Douglas Mawson died on 14 October 1958. Mawson's Legacy Mawson's inborn qualities as an explorer, his methodical approach to the job in hand, his tenacity and resolution and an ability to inspire and motivate others were revealed as much in his career as an academic geologist as in his exploits as an Antarctic explorer. He was essentially a man of action, a trail-blazer, single-minded in the pursuit of his goals, at times impatient for results, and who could, at times, be protective to the point of touchiness, of his own areas of interest. His inherited characteristics, hard-headed Yorkshire practicality, stubborness and reticence, with a streak of romanticism (from his father) were blended with a pioneer Australian belief in opportunity and self-reliance to produce a complex character of contrasting and at times conflicting elements. He was a man not easily assessed or categorised. Geology is arguably the most adventurous of all the field sciences (the first scientist to visit the moon was a geologist), and Mawson in his life and career, both in the Antarctic and in his geological investigations in outback South Australia, exemplified both the scientific-explorer and explorer-scientist. Mawson's scientific career spanned, almost exactly, the first half of the twentieth century. When he entered the University of Sydney in 1899 the scientific world-view was of a cooling, shrinking earth not more than one hundred million years old, the limits having been set by the physicists. Geologists, whose investigations had by then spread to all parts of the globe, intuitively sensed a more ancient earth but lacked the means to prove it. Mountain belts were seen as a crustal adjustment to a cooling interior and most geologists accepted the relative permanence of the continents and oceans. In retrospect, the years 1900-1950 can be viewed as a relatively quiet episode in the development of geological science, enlivened by two developments early in the century : the application of the discovery of radioactivity to the dating of rocks, and Wegener's hypothesis of continental drift. Each in its own time, the first rapidly, the second much more slowly, became a vital element in the future Plate Tectonic revolution. The recognition that the decay of radioactive minerals produced heat freed geologists from the shackles of a cooling earth and a restricted time-scale by providing a source of self-sustaining energy and a means of dating the crustal rocks. By 1914, the estimated age of the earth had been pushed back to over one thousand million years. Mawson's work on radioactive minerals in Australia as a young student and his later discoveries in South Australia gave him a pioneer role in this new development and he was always very proud of his work in this field (see Mawson 1944b). But in his day Australia was effectively too remote from the rest of the world for its small and largely inaccessible deposits to be economically viable, and its universities (certainly Adelaide) too poor to engage in expensive age-dating equipment. In his later years Mawson was keenly interested in these developments overseas and their future potential for dating the old rocks in Australia. Communication between scientists was severely curtailed during the two world wars which dominated the first half of the century, although technical developments during the second war were to help revolutionise geological and geophysical exploration both on land and in the ocean basins. Mawson's own career was savagely disrupted by the First World War and his more active years were over before the benefits of post- Second World War technical advances became widely felt. The inter-war years and those immediately after the second war were his most productive period for research, more particularly from the mid- 1930s when his active involvement in Antarctic exploration had ended. Although the general outline of South Australian geology was well known by this time, no systematic work had been attempted and for much of the State 134 D. W. CORBETT information remained highly generalised and lacking in detail. As previously indicated, Mawson's early researches were essentially exploratory reconnaissances, becoming more systematic with his work in the Flinders from the 1930s. His detailed descriptions of rock units and measured cross-sections of key sequences provided the basis for the regional mapping program to be instigated by the Geological Survey in the 1960s. Mawson's research was very much an integral part of the growth and development of his department. His field studies in the Flinders were carried out with small groups of advanced students as part of their course work, while in the Southeast he published in collaboration with post- graduate and former students as well as directing them in projects of their own. His teaching is remembered by past students as being of variable quality; he was at his best, according to Alderman, when extemporising on subjects that particularly interested him. There is no doubt that he could inspire his students and a considerable number went on to make distinguished careers in geology and mineralogy, among them E. A. Rudd, L. W. Parkin, R. C. Sprigg, W. B. Dallwitz, D. E. Thomas, E. R. Segnit, B. Skinner, R. L. Crocker, M. Reynolds, B. Daily, R. Sweatman, A. Kleeman, B. Forbes and G. Chinner. Mawson, the man of action, always seemed to have been fighting a lack of time to complete his projects; always planning another field trip and, one senses, chafing at the administrative burden placed on him as a highly responsible and consistently conscientious head of department. Never an armchair geologist, Mawson did not have the time or the inclination to become involved with the theoretical issues of the day, the most contentious of which was the idea of continental drift. Sir Mark Oliphant, a science student at the University in the 1920s, has recalled attending a debate on the subject between Mawson and Wood-Jones the anatomist, during which Mawson scathingly dismissed the concept of drift, citing the rigidity of the ocean floors and the lack of any known mechanism that could power the process. Sir Douglas Mawson made an outstanding contribution to South Australian geology. His list of 102 publications, the majority published in the Transactions of the Royal Society of South Australia, provide the material evidence of his work. He also left a strong department which he had established and which was developing rapidly in teaching and research. While it is idle to speculate, one cannot help but wonder what he might have accomplished as a geologist if so much of his time and energy had not been devoted to the Antarctic. He was not able to settle down to an ordered academic life and an on-going research program until he reached his fifties. From then on his geological output was sustained until the end of his career. It was a remarkable achievement. Acknowledgments The research on which this paper is based was begun when the author was a staff member in the Mawson Graduate Centre for Environmental Studies at the University of Adelaide and completed as an Honorary Research Associate at the South Australian Museum. It relies heavily on a study of the Mawson Papers housed in the Special Collections archive of the Barr Smith Library, The University of Adelaide, and Mawson's published papers. Research was also carried out on the Edgeworth David papers held by the Fisher Library at the University of Sydney and in the library of the Scott Polar Research Institute, Cambridge. For valuable assistance during the course of this work, thanks are due to: Susan Woodbum and the staff of Special Collections, The University of Adelaide; Mr Ken Smith, Archivist, The Fisher Library and Professor David Branagan, The University of Sydney; Mr Robert Headland, Curator and Archivist, The Scott Polar Research Institute, Cambridge. For on-going encouragement and support for this research and for much helpful discussion on matters relating to Mawson, grateful thanks are extended to Mr Richard Ferguson, Polar historian; Emeritus Professor E. A. Rudd, The University of Adelaide; Dr Victor Gostm of the Geology Department, The University of Adelaide and Dr Allan Pring, Curator of Minerals, The South Australian Museum. Particular thanks are due to Dr Robin Oliver of the Geology Department, University of Adelaide, for his critical review of the original draft of this paper, and his many useful suggestions. Selected Bibliography of Douglas Mawson (See Alderman (1959) for a Memoir and Bibliography of Mawson. Innes and Duff (1990) give a complete list of his scientific and other papers together with biographical works on Mawson). 903 (with T. G. TAYLOR). The geology of Mittagong. Journal and Proceedings of the Royal Society of New South Wales 37: 306-350." DOUGLAS MAWSON GEOLOGIST 135 1904 (with T. H. LABY). Preliminary observations on radioactivity and the occurrence of radium in Australian minerals. Journal and Proceedings of the Royal Society of New South Wales 38: 382-389. 1905 The geology of the New Hebrides. Proceedings of the Linnean Society of New South Wales 3: 400-485. 1906 On certain new mineral species associated with carnotite in the radioactive orebody near Olary. Transactions of the Royal Society' of South Australia 30: 188-193. 1912a Geological investigations in the Broken Hill area. Memoir of the Royal Society of South Australia 2(4): 211-319. 1 9 1 2b PreCambrian areas in the north-eastern portion of South Australia and the Barrier, New South Wales. Reports of the Australasian Association for the Advancement of Science 13: 188-191. 1915 The Home of the Blizzard. Heinemann, London. [Extracts quoted are taken from the abridged popular edition (Hodder - Stoughton 1930), published as a facsimile edition by the Wakefield Press, 1996.] 1923 Igneous rocks of the Mount Painter belt. Transactions of the Royal Society of South Australia 47; 37^387. 1925 Evidence and indications of algal contributions in the Cambrian and Pre-Cambrian limestones of South Australia. Transactions of the Royal Society of South Australia 49: 186-190. 1926a Wooltana basic igneous belt. Transactions of the Royal Society of South Australia 50: 192-200. 1926b A brief resume of present knowledge relating to the igneous rocks of South Australia. Reports of the Australasian Association for the Advancement of Science 18: 229-274 1927a The Paralana hot spring. Transactions of the Royal Society of South Australia 51: 391-397. 1927b Geological notes on the area along the north- eastern margin of the north-eastern portion of the Willouran Range. Transactions of the Royal Society of South Australia 51 : 386-390. 1930 (with C. T. MADIGAN). The pre-Ordovician rocks of the McDonnell Ranges, Central Australia. Quarterly Journal of the Geological Society of LondonS6: 415^28. 1938 The Mount Carnarvon Series of Proterozoic Age. Transactions of the Royal Societ\> of South Australia 62:347-351. 1939a The Cambrian sequence in the Wirrealpa Basin. Transactions of the Royal Society of South Australia 63(2): 331-347. 1 939b The late Proterozoic sediments of South Australia. Reports of the Australasian and New Zealand Association for the Advancement of Science 24: 80- 88. 1942 The structural character of the Flinders Ranges. Transactions of the Royal Society of South Australia 66(2): 262-272. 1944a Unpublished memorandum on housing the department of geology, The University of Adelaide. February 1944. 1944b The nature and occurrence of uraniferous mineral deposits in South Australia. Transactions of the Royal Society of South Australia 68(2): 334-357. 1945a (with W. B. DALLWITZ). The soda-rich leucogranite cupolas of Umberatana. Transactions of the Royal Society of South Australia 69(1): 22- 49. 1945b (with W. B. DALLWITZ). Scapolitised dolomites of Yankaninna. Transactions of the Royal Society of South Australia 69(2): 212-216. 1947 The Adelaide Series as developed along the western margin of the Flinders Ranges. Transactions of the Royal Society of South Australia 71(2): 259- 280. 1949a The Elatina glaciation: a third recurrence of glaciation evidenced in the Adelaide System. Transactions of the Royal Society of South Australia 73: 117-121. 1949b The late Precambrian ice-age and glacial record of the Bibliando Dome. Journal and Proceedings of the Royal Society of New South Wales 82: 150- 174. 1950 (with R. C. SPRIGG). Subdivision of the Adelaide System. Australian Journal of Science 13(3): 69- 70. 1958 Australites in the vicinity of Florieton, South Australia. Transactions of the Royal Society of South Australia HI: 161-163. References ALDERMAN, A. R. 1959. Professor Sir Douglas Mawson Memoir and Bibliography. Transactions of the Royal Society of South Australia 82: 1-6. ALDERMAN, A. R. 1967. The development of geology in South Australia. Records of the Australian Academy of Science 1(2): 30 52. BICKEL, L. 1977. This Accursed Land. Macmillan Australia. CORBETT, D. W., COOPER, B. J. & MOONEY, P. M. 1986. 'Geology' in 'Ideas and Endeavours - the Natural Sciences in South Australia'. Royal Society of South Australia Occasional Publication Number 5. DAVID, T. W. E. (with R. HELMS & E. F. PITTMAN) 1901. Geological notes on Kosciusko, with special reference to evidences of glacial action. Proceedings of the Linnean Society of New South Wales 26: 26 74. 136 D. W. CORBETT DAVID, T. W. E. & PRIESTLEY, R. 1914. Geology Vol. 1. Reports, on the Scientific Investigations, the British Antarctic Expedition, 1907-08. DAVID, T. W. E. 1928. Notes on the newly-discovered fossils in the Adelaide Series (Lipalian?), South Australia. Transactions of the Royal Society of South Australia SI: 191-209. DAVID, T. W. E. & TILL YARD, R. J. 1936. Memoir on fossils of the late Pre-Cambrian (Newer Proterozoic) from the Adelaide Series, South Australia. Angus and Robertson, Sydney. FERRAR, H. T. 1907. Report on the field geology of the region explored during the 'Discovery' Antarctic Expedition, 1901-04. National Antarctic Expedition 1901-04, Natural History, 1, Geology (Field Geology: Petrography). British Museum, London. GLAESSNER, M. F. & DAILY, B. 1959. The geology of the late Precambrian fauna of the Ediacara Fossil Reserve. Records of the South Australian Museum 13(3): 369-401. HOWCHIN, W. 1897. On the occurrence of Lower Cambrian fossils in the Mount Lofty Ranges. Transactions of the Royal Society of South Australia 20(2): 74-86. HOWCHIN, W. 1901. Preliminary note on the existence of glacial beds of Cambrian age in South Australia. Transactions of the Royal Society of South Australia 25: 10-13. HOWCHIN, W. 1914. The occurrence of the Genus Cryptozoon in the ? Cambrian of Australia. Transactions of the Royal Society of South Australia 38: 1-10. HOWCHIN, W. 1922. A geological traverse of the Flinders Range from the Parachilna Gorge to the Lake Frome Plains. Transactions of the Royal Society of South Australia 46: 46-82. INNES, M. & DUFF, H. 1990. 'Mawson's Papers 1 . The Mawson Institute for Antarctic Research, The University of Adelaide. JACKA, F. 1986. Mawson, Sir Douglas(1882-1958). Australian Dictionary of Biography 10: 454-457. MAWSON, P. 1962. 'Mawson of the Antarctic'. Longmans, London. PARER, P. & PARER-COOK, E. 1983. 'Douglas Mawson the Explorer'. Alella Books ABC. SEGNIT, R. W. 1939. 'The Pre-Cambrian Cambrian Succession'. Bulletin of the Geological Survey of South Australia 18. SHACKLETON, E. 1911. 'The Heart of the Antarctic'. Heinemann, London. SPRIGG, R. C. 1947. Early Cambrian (?) jellyfishes from the Flinders Ranges, South Australia. Transactions of the Royal Society of South Australia 72(2): 212-224. SPRIGG, R. C. 1949. Early Cambrian 'Jellyfishes' of Ediacara, South Australia. Transactions of the Royal Society of South Australia 73: 72-99. SPRIGG, R. C. 1986. The Adelaide Geosyncline: a century of controversy. Earth Sciences History 5(1): 66-83. REVISION OF AUSTRALIAN ENOCHRUS THOMSON (COLEOPTERA: HYDROPHILIDAE) C.H.S. WATTS Summary The Australian members of the hydrophilid genus Enochrus are revised and redescribed. A key to the fourteen species recognised is given. Six species are described as new: E. aliciae, E. eubenangeei, E. isabellae, E. pseudoweiri, E. samae, and E. weiri. The following synonymies are proposed: E. mjobergi Knisch, 1921 - E. deserticola (Blackburn, 1896); E. andersoni (Blackburn, 1896) = E. eyrensis (Blackburn, 1894); E. persimilis Regimbart, 1908 = E. eyrensis (Blackburn, 1894); E. pullus (Fauvel, 1883) = E. esuriens (Walker, 1858) and E. artensis (Fauvel, 1883) = E. maculiceps (MacLeay, 1873). The New Caledonian species E. caledonicus (Fauvel, 1883) is considered a synonym of E. elongatus (MacLeay, 1873). REVISION OF AUSTRALIAN ENOCHRUS THOMSON (COLEOPTERA: HYDROPHILIDAE) C. H. S. WATTS WATTS, C. H. S. 1998. Revision of Australian Enochrus Thomson (Coleoptera; Hydrophilidae). Records of the South Australian Museum 30(2): 137-156. The Australian members of the hydrophilid genus Enochrus are revised and redescribed. A key to the fourteen species recognised is given. Six species are described as new: E. aliciae, E. eubenangeei, E. isabellae, E. pseudoweiri, E. samae, and E. weiri. The following synonymies are proposed: E. mjobergi Knisch, 1921 - E. deserticola (Blackburn, 1896); E. andersoni (Blackburn, 1896) - E. eyrensis (Blackburn, 1894); E. persimilis Regimbart, 1908 * E. eyrensis (Blackburn, 1894); E. pullus (Fauvel, 1883) = E. esuriens (Walker, 1858) and E. artensis (Fauvel, 1883) = E. maculiceps (MacLeay, 1873). The New Caledonian species E. caledonicus (Fauvel, 1883) is considered a synonym of E. elongatus (MacLeay, 1873). C. H. S. Watts, South Australian Museum, North Terrace, Adelaide, South Australia 5000. Manuscript received 19 March 1997. The hydrophilid genus Enochrus Thomson, 1859 is world wide in distribution (Hansen 1991). In Australia its members are the commonest and most widespread hydrophilids and occur in virtually all fresh water bodies from small stagnant pools to the banks of major rivers and lakes. Despite this prominence, their taxonomy is so bad that no species can be identified with any confidence. Thirteen species have been named and several others have been recorded as being present but until now no revision has been attempted. Australian Enochrus are structurally very similar and difficult to separate. Apart from the aedeagi the only characters that I have found useful are size, the punctation of the upper surface, the slope of the mesosternal keel, and the colour pattern of the head. The colour pattern of the head is often a reasonable indicator of species, whereas elytral and pronotal colour can vary from predominantly black to totally light reddish in the same species. Within Enochrus six subgenera are recognised (Hansen 1990). Three of these occur in Australia: Hydatotrephis MacLeay, 1873 with one species; Enochrus Thomson, 1859 with one species; and Methydrus Rey, 1885 with 12 species. At first sight E. (H.) mastersi is very similar to the distantly related Limnoxenus zealandicus Broun and is often confused with it in collections. Both are moderate sized (10-15 mm), shiny black and relatively common, particularly L. zealandicus. A closer inspection readily separates them: L. zealandicus has well defined punctate elytral striae and swimming hairs on the meso- and meta-tibiae, both of which are absent in E, (H.) mastersi. Characters of the maxillary palpi and mesosternum also separate the two genera (Hansen 1990). The one Australian Enochrus (Enochrus) species, E. (E.) peregrinus Knisch, 1922, is small and in general appearance resembles several Australian Enochrus (Methydrus) species. It is known from only one specimen labelled from Sydney, although there were three in the original collection. It may eventually prove to have been mis-labelled and not Australian. The remaining 12 Australian species are all in the subgenus Methydrus. They are usually small (2 8 mm), oval, rather flat species, shiny with light-testaceous to black colouring with at least the area immediately in front of the eyes yellowish. Two species are larger and predominantly black, superficially resembling E. (H.) mastersi. Chasmogenus nitescens Fauvel resembles some Enochrus (Methydrus) species and is often confused with it in collections. Its small size (< 5 mm) and lack of yellowish areas in front of its eyes readily separate it from similar sized Enochrus - as well as many structural characters mentioned later under 'Systematics'. One additional species, Philhydrus marmoratus W. MacLeay, 1873, has been included in Enochrus in the past but has been shown by Gentili, 1981 to belong in Laccobius. For a full discussion of synonyms and subgenera of Enochrus see Hansen, 1991. Despite their being ubiquitous in freshwater the biology of Australian Enochrus is poorly known. 38 C. H. S. WATTS Anderson (1976) described the larvae of E. (M.) maculiceps and E. (M.) elongatus and gave brief notes on their habitat and breeding. All species come readily to light and are often collected in large numbers in light traps, particularly in northern areas during the wet season from December to February. I know of no study of the ecology of any Australian species other than inclusion in general faunal lists of invertebrates collected in various water bodies. The collections from which specimens were examined are listed under the following abbreviations: AM Australian Museum, Sydney AN1C Australian National Insect Collection BM(NH) Natural History Museum, London IRSNB Institut Royal des Sciences Naturelles de Belgique, Bruxelles MNHN Museum National d'Histoire Naturelle, Paris NMV Museum of Victoria NRS Naturhistoriska Riksmuseet, Stockholm NTM Northern Territory Museum QDPIM Queensland Department of Primary Industries, Mareeba QM Queensland Museum, Brisbane SAMA South Australian Museum, Adelaide UQIC University of Queensland Insect Collection, Brisbane WAM Western Australian Museum, Perth. Systematics Australian members of the genus Enochrus can be separated from other Australian Hydrophilids by the following combination of characters. Length 2-9 mm, oval, more flattened than convex; second segment of maxillary palpi more or less curved outwards, apical segment slightly asymmetrical with straighter inner face; without contiguous ventral keel but with variable keel on mesosternum; without swimming hairs on tibiae; without strongly marked striae on elytron; with systematic punctures (rows or fields of coarse setiferous punctures usually distinctly larger than others) on head and pronotum (masked in some E. malabarensis); in all but E. mastersi and E. aliciae at least the area immediately in front of eyes lighter (see also Hansen 1991). The subgenus Hydatotrephis MacLeay can be separated from Australian members of the subgenera Enochrus Thomson and Methydrus Rey as follows: mesosternal protuberance triangular, cone-shaped, abruptly truncated posteriorly which does not reach the front edge of mesocoxae (Fig. 19), sides of mesosternum evenly converging anteriorly, apical and penultimate segments of maxillary palpi subequal. Enochrus (Methydrus) have a more keel-like mesosternal process usually extending backwards between the mesocoxae (Figs 15-18; 20), the apical segment of the maxillary palpi shorter than penultimate and sides of mesosternum subparallel to strongly convex anteriorly. Apart from E. (M.) aliciae and E. (M.) eubenangeei, the Australian members of E. (Methydrus) are smaller (< 8 mm), often testaceous in colour, and have at least a lighter patch of colour in front of the eyes. The aedeagus of E. (H.) mastersi differs from that of other Australian Enochrus (Fig. 14; Hansen 1990). In these there is a variably shaped basal ventral plate beyond which projects a thinner apical portion in all but E. (M.) aliciae which lacks this apical portion (Fig. 11). In E. (H.) mastersi the aedeagus is of more uniform construction with a smooth dorsal surface and the ventral surface consists of a broad open groove formed by infolding of the two sides. There is one species, E. (E.) peregrinus Knisch, of the subgenus Enochrus in Australia. It is similar in looks to the smaller Methydrus species but is readily separated from them by the short stout maxillary palpi with approximately equal sized apical two segments and the diverging elytral suture lines. By utilising the male genitalia species can be differentiated. However without these I have found it impossible to reliably separate all species, in particular E. weiri and E. pseudoweiri and the lighter colour phase of E. deserticola and E. maculipes. This is reflected in the keys and descriptions that follow. 1 have restricted my descriptions to the few characters that serve to differentiate the species and I have provided two keys, one utilising all available characters, the other characters other than the male genitalia. Key to Australian Enochrus Thomson 1859, not using characters of the male genitalia 1 . — Maxillary palpi stout, first segment shorter than distance from front edge of eye to front of clypeus, apical segment subequal in length to penultimate 2 Maxillary palpi more elongate, first segment AUSTRALIAN ENOCHRUS 139 longer than distance from front edge of eye to front margin of clypeus, apical segment about 2/3 length of penultimate (subgenus Methydrus) 3 2. — > 6.0 mm. Black. Sutural lines on elytra subparallel in anterior quarter , E. {Hydatotrephis) mastersi (MacLeay) — < 5.0 mm. Dark testaceous. Sutural lines on elytra diverging noticeably in anterior quarter .. E. (Enochrus) peregrinus Knisch 3. — Large (>6 mm). Black (pronotum may be reddish at edges) 4 Small (< 6 mm). With at least the areas in front of eyes reddish-yellow 5 4. - Punctures towards sides of elytra well impressed, not much smaller than others E. (M.) euhenangeei sp. nov. — Punctures towards sides of elytra subobsolete, much smaller than those on disc E. (M.) aliciae sp. nov. 5. — Dorsal surface reddish-yellow, rear of head at most somewhat darker. Elytra moderately punctate, length > 4.0 mm, pro-, meso- and metaclaws modified in male E. (M.) elongatus (W . MacLeay) - Dorsal surface reddish-yellow to black, rear of head dark and much of rest of head in many species, males with only pro-claws modified 6 6. — Dorsal surface strongly punctured, punctures on elytra as strong or stronger than those on head, many punctures on elytra >half diameter of serial punctures which are often hard to trace. Front of head 1/3 or more black 7 — Dorsal surface weakly punctured, punctures on elytra tend to be weaker than those on head, punctures on elytra < half size of serial punctures which are usually distinct 9 1, — Length >3.5 mm. Black morphs present. Maxillary palpi with dark tips 8 Length < 4.0 mm. Black morphs not known. Maxillary palpi usually pale throughout E. (E.) malabarensis Regimbart 8. — Thinner outer portion of mesosternal keel well developed (Fig. 16). Front of head predominantly black E. (M.) eyrensis (Blackburn) — Thinner outer portion of mesosternal keel less developed (Fig. 15). Amount of black on front of head variable E. (M.) samae sp. nov. 9. — Length 2.2 mm - 2.6 mm. Mesosternal keel weakly developed (Fig. 1 8). Front of head > 2/3 black E. (M.) esuriens (Walker) — Length > 2.6 mm. Mesosternal keel well developed (Fig. 17) 10 10. — Larger (4.5-6 mm). Black, except region in front of eyes light. Elytral punctures small but usually well marked E. (M.) isabellae sp. nov. — Smaller (3-4.5 mm) E. (M.) deserticola complex (£. (M.) deserticola (Blackburn), E. (M.) maculiceps (W. MacLeay), E. (M.) weiri sp. nov and E. (M.) pseudoweiri sp. nov). (E. (M.) deserticola only species with black morphs and individuals > 4.2 mm long. Enochrus weiri and E. pseudoweiri always have pale palpi.) Key to Australian Enochrus incorporating characters of the male glnitalia — Maxillary palpi stout, pseudo first segment shorter than distance from front edge of eye to front of clypeus, apical and penultimate segments subequal in length 2 — Maxillary palpi more elongate, pseudo first segment longer than distance from front edge of eye to front margin of clypeus, apical segment noticeably shorter than penultimate (subgenus Methydrus) 3 — > 6.0 mm. Black. Sutural lines on elytra subparallel in anterior quarter E. (Hydatotrephis) mastersi (MacLeay) < 5.0 mm. Dark testaceous. Sutural lines on elytra noticeably diverging in anterior quarter .. E. (Enochrus) peregrinus Knisch — Large (> 6.0 mm). Black (pronotum may be reddish at edges) 4 — Small (< 6.0 mm). With at least the areas in front of eyes reddish-yellow 5 — Aedeagus broad reaching little beyond half length of parameres (Fig. 1 1). Parameres flat. .. E. (M.) aliciae sp. nov. Aedeagus with broad basal piece and long thin apical piece bending upwards and reaching nearly to end of parameres, which also bend upwards (Fig. 10) E. (M.) euhenangeei sp. nov. — Tips of parameres hooked (e.g. Fig. 4) 6 — Tips of parameres not hooked (e.g. Fig. 1) 9 — Dorsal surface reddish-yellow, rear of head at most somewhat darker. Aedeagus tip strongly crotchet-shaped (Fig. 7). Pro-, meso- and metaclaws modified in male E. (M.) elongatus (W. MacLeay) Variably coloured but always with rear of head clearly darker than at least portions of front of head. Aedeagus with little or no dorsal/ventral expansion at tip, males with only proclaws modified 7 140 C. H. S. WATTS 7. — Dorsal surface strongly punctured, many punctures on elytra > half size of serial punctures E. (M.) eyrensis (Blackburn) — Dorsal surface weakly punctured, punctures on elytra < half size of serial punctures 8 8. — Larger (4.5 - 6.0 mm). Black, except in front of eyes. Aedeagus shorter and apical portion broader (Fig. 6) E. (M.) isabellae sp, nov. — Smaller (3.0-4.0 mm). Pale morphs common. Aedeagus longer and apical portion thinner (Fig. 4) E. (M.) deserticola (Blackburn) 9. — Length <2.8 mm. Front of head > 2/3 black. Mesosternal keel weak (Fig. 18) E. (M.) esuriens (Walker) — Not as above 10 1 0. — Dorsal surface strongly punctured, punctures on elytra as strong as, or stronger than, those on head, many punctures on elytra > half diameter of serial punctures 1 1 — Dorsal surface weakly punctured, punctures on elytra tend to be weaker than those on head, punctures on elytra < half size of serial punctures 12 11. — Larger (4.4-5.6 mm). Black morphs present. Aedeagus with moderate apical pad, inner edges of parameres sinuate (Fig. 9) E. (M.) samae sp. nov. — Smaller (<3.8 mm). No black morphs. Aedeagus narrow with, at most, a weak pad, inner edges of parameres straight (Fig. 12) E. (M.) malabarensis Regimbart 12. — Tips of parameres truncated or weakly bifid (Fig. 8) E. (M.) weir i sp. nov. — Tips of parameres rounded (Figs I, 2, 3) 13 13. — Aedeagus moderately broad with slight to well- marked apical pad (Figs 1, 2) E. (M.) maculiceps (W. MacLeay) — Aedeagus narrow, sharply pointed, lacking apical pad (Fig. 3) E. (M.) pseudoweiri sp. nov. Descriptions Subgenus Enochrus Enochrus (Enochrus) peregrinus Knisch, 1922 Types Lectotype: 9 'Sidney Mus. Godeffroy No. 10705, coll. Knisch, coll. d'Orchymont'; 'ex. col. Knisch, No. 5251438, coll. d'Orchymont'; 'Knisch det. 1921, Enochrus peregrinus m'; 'Coll. A. Knisch, TYPUS', on red card, in IRSNB, herein designated. Knisch (1922) mentioned three specimens (No. 10705) in the Hamburg Zoological Museum. I have been able to trace only one. He mentioned that one was labelled 'M. Regimbart det. 1905' and the others with 'Wehncke determ'. The surviving specimen, now in Brussels, lacks such a label but would otherwise appear to be part of the syntype series. Description (number examined, 1) Length 3.2 mm. Broadly oval. Elytra dark- testaceous, pronotum very slightly lighter, head black except for small indistinct testaceous patch along margin in front of each eye, maxillary palpi light-testaceous, tips darker. Punctures on head moderate, approaching size of eye facet, most separated from each other by at least puncture width. Systematic punctures rather sparse, about twice diameter of adjacent punctures. Pronotum rather weakly and sparsely punctured, serial punctures also relatively small, indistinct, about 2x diameter of adjacent punctures, punctures on elytra stronger than on pronotum, sparse, many 2x diameter or more apart, a little stronger laterally. Serial punctures not traceable. Maxillary palpi short, stout, pseudo basal segment not reaching front edge of eye, apical and penultimate segment subequal in size. Male: Unknown. Distribution Known only from Sydney, the type locality, Remarks This unique specimen is separated from all other Australian Enochrus, other than the much larger E. mastersi, by the short stout maxillary palpi. The masking of the elytral strial punctures is found only in the much larger E, eyrensis and E. samae and in E. malabarensis. Enochrus malabarensis differs from E. peregrinus by the more normal (for Australian species) maxillary palpi, the more extensive testaceous areas in front of the eyes and lighter coloured elytra. The elytral sutures continue to diverge noticeably in anterior quarter whereas in E. malabarensis they are subparallel. Most E. malabarensis also have stronger punctures, but this character is variable and tends to be relatively weak in the few Sydney area specimens I have seen. The pronotal punctures in E. peregrinus are weaker than those on elytra whereas in E. malabarensis they are similar sized. AUSTRALIAN ENOCHRUS 141 The fact that only the type specimen is known worries me. For all the other small Australian Enochrus, specimens are relatively abundant. Also it is the only known specimen in the Australian region of the subgenus Enochrus (Hansen 1990). This is the sort of situation where mislabelling should be considered. This is beyond the scope of this revision. Subgenus Hydatotrephis Enochrus (Hydatotrephis) mastersi (MacLeay) Hydatotrephis mastersi MacLeay, 1871 = Parana simplex Knisch, 1922: Hansen, 1990 Enochrus (Hydatotrephis) mastersi (MacLeay): Hansen, 1990 Types Hydatotrephis mastersi: Lectotype: ( 9- ), (round red label), 'K 1950T, "Hydatotrephis mastersi Mc L.W. Gayndah'. 'Lectotype: Hydatotrephis mastersi W. J. MacLeay designated by M. Hansen 1990', AM. Parana simplex: Lectotype: (unsexed) 'Sidney Mus. Godeffroy. No. 10704'; 'Philhydrus x M. Regimbart determ. 1905'; 'Parana simplex Knisch n.g. et sp. A. Knisch det. 1921', round black label; 'Lectotype Parana simplex Knisch designated by M. Hansen 1988', IRSNB. Synonymy after Hansen 1990. Description (number examined, 71) Figs 14, 19 Size 8.0 mm - 9.1 mm. Broadly oval. Relatively flat. Black; lateral edges of elytra, pronotum and head, palpi and apical portions of legs diffusely dark-testaceous. Maxillary palpi stout, length about width of head in front of eyes, apical and penultimate segments subcqual in length, together a tittle larger than pseudo first segment. Head strongly punctured, each about size of eye facet, most closer than a puncture width apart, systematic punctures 2-3x size of normal punctures. Pronotum with punctures a little smaller, shallower and more separated than on head, systematic punctures well marked, with sharp groove sometimes weak or lacking on front edge. Elytra punctured as on pronotum, each elytron with four rows of larger serial punctures, punctures in rows increasing in both number and scatter towards sides, elytra weakly grooved and flanged laterally, a few scattered large punctures between puncture rows two and three counting from suture. Mesosternal keel cone-like, strongly raised, narrowly triangular from both lateral and front views, sharply truncated behind just prior to mesocoxal cavities so that no part of the keel extends between mesocoxae. Male: Aedcagus broad, lacking collar or ventral plate, sides folded in ventrally tending to form broad channel. Parameres stout, broad, narrowing shortly before tips which tend to be turned inward to varying degrees. Distribution Australian Capital Territory Black Mt, ANIC; Canberra, SAMA, ANIC; 3 km E Piccadilly Circus, ANIC; 6 km NE Piccadilly Circus, ANIC. New South Wales Allyn River, ANIC; Barrington River, ANIC; Blue Mts, ANIC Chichester St Forrest, ANIC; Coffs Harbour UQIC; Como, ANIC; Galston, SAMA; Lansdowne via Taree, ANIC; Muswellbrook, ANIC; Nerringa, SAMA; 12 km NW Nellinger, ANIC; 2km NE by N Rousmill, ANIC; Salisbury UQIC; Stanwell Park UQIC; Sydney, SAMA, ANIC; Tooloom Plateau via Woodenbong UQIC; Ulladulla, ANIC; Valery, ANIC; Wee Jasper, ANIC. Northern Territory 30 ml W Alice Springs, SAMA; Stanley Chasm, SAMA. Queensland Brisbane UQIC; Bulburin State Forest via Many Peaks UQIC; Bunya Mt, ANIC; Carnarvon UQIC; Discot, ANIC; 13 km SW by S Gordonvale, ANIC; Joalah Nt Pk Tamborine Mt, ANIC; Mt Norman, via Wallangarra UQIC; Mt Spec, ANIC; Upper Cedar Ck, via Sandford UQIC; 8 km E Wallaman Falls, SAMA. Victoria Bagots Ck, SAMA; Ferntree Gully UQIC; Healesville, SAMA; 10 mis N of Valencia Creek via Maffra UQIC; Victorian Alps, SAMA. Remarks A widespread and relatively common species which, within Australian Enochrus, can only be confused with E. aliciae and E. eubenangeei. It is readily separated from these species by the stout maxillary palpi which have the last two segments subequal instead of having the apical segment about two-thirds the length of the penultimate as 142 C. H. S. WATTS in all other Australian Enochrus. The cone-like shape of the mesosternal keel is also diagnostic. Enochrus mastersi is most frequently confused with a very different, but similar looking and very common species, Limnoxenus zealandicus, of the subtribe Hydrobiina. Although very similar in general appearance Limnoxenus can be separated from E. mastersi by the presence of nine well- marked series of punctures on each elytron, swimming hairs on the meso- and metatarsi and virtual absence of rugose portions on meso- and metafemora. Subgenus Methydrus Enochrus (Methydrus) aliciae sp. nov. Types Holotype: o\ '15°16*S, 144°59'E, 14 km W by N of Hope Vaie Mission, Qld, 8-10 Oct. 1980, T. Weir', ANIC. Paratypes: 13, same data as holotype, 10 ANIC, 3 SAMA; 1, '12*31% 132°54'E, 9 km N by E of Mudginbarry Hs„ N.T., 10.vi.1973, T. Weir and A. Allwood, M3985", NTM. Description (number examined, 17) Figs 11, 20 Length 7.5 mm - 8.2 mm. Broadly oval, relatively flat. Shiny, black, appendages, apical and lateral edges of head, pronotum and elytra dark-reddish. Maxillary palpi moderately long, pseudo first segment longer than distance from eye to front of clypeus, longer than maxillary stipe, apical segment 2 I to 3 / 4 length of penultimate. Head strongly punctate, punctures about puncture width apart, about size of eye facet or a little smaller. Systematic punctures large about 3x size of adjacent punctures. Punctures on pronotum similar to or a little weaker than those on head, systematic punctures well marked, 3~4x diameter of adjacent punctures, rear and lateral margins weakly grooved and flanged, front margin grooved to about level of inner edge of eyes. Elytral punctures as on pronotum, more weakly impressed towards rear. Serial punctures in four loose rows, sparse and indistinct, more numerous but more scattered laterally, elytra weakly grooved and flanged laterally. Mesosternal pillar well marked, triangular in both front and lateral views, posterior edge often steeper near top of keel, reaching a short distance between mesocoxae. Male: Aedeagus very short, broad, rounded, lacking narrower apical section. Parameres squat, inner edge a little sinuate, tips obliquely truncated on inside or curved weakly outwards. Proclaw straightened with strong basal swelling, mesoclaw thickened, curved with basal swelling, metaclaw straightened and greatly swollen in basal half. In female claws similar but with straightening and basal swelling much weaker. Distribution Known only from the type localities in North Queensland and the Northern Territory. Remarks Readily separated from all other Australian Enochrus (Methydrus), other than E. eubenangeei, by its large size and almost all-black colouring. The very short, broad aedeagus, and weak lateral punctures on the elytra separate it from the otherwise very similar E. eubenangeei. In general appearance it is similar to E. (H.) mastersi but readily separated by its longer maxillary palpi with the apical segment shorter than penultimate. See also under remarks in E. mastersi. Enochrus (Methydrus) deserticola (Blackburn) Philhydrus deserticola Blackburn, 1896 Enochrus (Lumetus) deserticola (Blackburn): Knischl924 —Enochrus mjobergi Knisch, 1921: syn. nov. —Enochrus (Lumetus) mjobergi Knisch: Knisch 1924 ^Philhydrus temporalis Regimbart, 1908: syn. nov. ^Enochrus (Lumetus) temporalis (Regimbart): Knisch 1924 Types Enochrus deserticola: Lectotype: '5487, Bl. Paisley T' top specimen of two separately carded on same pin, BM(NH), herein designated. Paralectotypes: 1 unlabelled but mounted below lectotype on same pin, BM(NH); 1 6 '5487 Palm Cr', ' Philhydrus deserticola Blackb. Co- type', SAMA; 3, 'Reedy Hole', NMV; 2, 'Ellery Ck\ NMV; 2, 'Paisley Bluff, NMV. Enochrus mjobergi: Lectotype: 6 'Cape York Penins'; 'Queensl. Mjoberg'; 'Type', ''Enochrus (Lum.) mjobergi m. n. sp., A. Knisch 1921 '; 'Typus', on red label NRS, herein designated. Paralectotypes: 1, 'Queensl. Mjoberg Cape York Penins'; 'Sjostedt don 1921'; 'Ex Coll. Knisch No. 5521544 Coll. d'Orchym'; 'Knis det AUSTRALIAN ENOCHRUS 43 Enochrus (L.) mjobergi m'; Coll. A. Knisch TYPUS', on red card. IRSNB; 3, 'Cape York Penins'; 'Queensl. Mjoberg', NRS. Synonymy after examination of types. Enochrus temporalis: Holotype: 'Avon R'; 'eaudouce'; ' Philhydrus temporalis, Reg. TYPUS % MNHN. Synonymy based on descriptions and examination of types (of E. temporalis in 1964, not currently available). Description (number examined, >1000) Fig. 4 Oblong oval, length 3.3 mm - 4.9 mm. Two colour morphs; one with dorsal surface black except light portions in front of eyes, other with elytra and pronotum light testaceous and head dark testaceous to black except for light testaceous area forward from eyes and a little wider than eyes, leaving central darker panel of variable width on front of head, underside dark testaceous, appendages lighter towards extremities except for maxillary palpi which usually have apical portion of last segment darker. Head covered with small, sharply impressed, well separated punctures a lot smaller than eye facets, a group of large punctures inwards from each eye, at least 4x diameter of other punctures. Punctures on pronotum as on head, but more shallowly impressed, systematic punctures as in E. elongatus, four or more times diameter of adjacent punctures. Elytral punctures increase somewhat in strength apically and laterally, serial punctures large, sparse and quite distinct at apex, interstrial punctures varying from weak to almost obsolete. Mesosternal keel well developed, front edge nearly perpendicular, ventral edge straight or slightly convex. Male: Aedeagus pointed, with weak dorsal pad at tip, collar a bit closer to tip than base. Parameres outwardly hooked at tips. Proclaw with basal lobe and rest of claw somewhat straightened. Meso- and mctaclaws, with slight basal lobes, weakly curved. Female: Claws only slightly weaker than males. Distribution New South Wales Bonville, ANIC; Chichester SF, ANIC Clarence River, NMV; Coffs Harbour, ANIC Fowlers Gap, ANIC; 10 km N Jabrru, QDPIM Lord Howe Island, SAMA; Moree, SAMA Valcry, ANIC; Wootton, ANIC. Northern Territory Adelaide River, ANIC; E Alligator River, AM; Bessie Spring, ANIC; Berry Spring, ANIC; 45 km SW by S Borroloola, ANIC; 7 km NW by N Mt Cahill Crossing East Alligator River, ANIC; 8 km N Mt Cahill, SAMA; 12 km NNW Mt Cahill, SAMA; 14 km S by W Cape Crawford, ANIC; Cooper Creek, ANIC; Daly River Mission, ANIC; 170 km E Daly Waters, SAMA; Darwin, ANIC, SAMA; Ellery Ck, NMV; 15 km S Elliot, SAMA; Harst Bluff, NMV; 12 km NE Humpty Doo, QDPIM; Jim Jim Creek, ANIC; Koogarra, ANIC; 4 km N McArthur R Stn, SAMA; 19 km SSE Mataranka, SAMA; 5 km E Mataranka, SAMA; Nourlangie Creek, ANIC; 18 km E by N Oenpelli, ANIC; Paisley Bluff, NMV; Pine Creek, QDPIM; Reedy Hole, NMV; 46 km S Renner Springs, SAMA; Stanley Chasm, SAMA, ANIC; 100 km W Tennant Creek, QDPIM; Tindal, ANIC; Wessel Island, ANIC. Queensland Ashmore, QM; Atherton, ANIC, QDPIM; 31 km NE Aramac, SAMA; Bakerville, QDPIM; Bamaga, SAMA; Brisbane, QM, UQIC; Benaraby, ANIC; 5 mis N Bloomfield River, ANIC; Bundaberg, ANIC; Bunya Mts, ANIC; Burketown, QM; Cairns, ANIC, SAMA; Calliope River, ANIC; Camooweal, QDPIM; Cape Tribulation, ANIC; Cardstone, ANIC; Chillagoe Creek, ANIC, QDPIM; Coen, QDPIM; 60 km S Coen, SAMA; Mt Coolum, ANIC; Cooktown, ANIC; 30 mis N Cooktown, UQIC; Crystal Creek, ANIC; Cunningham's Gap, ANIC; Dalhunty River, SAMA; Edungalba, ANIC; 5 km NE Edungalba, SAMA; Eidsvold, ANIC; Emu Park, UQIC; Fletcher, QM; Gayndah, UQIC; Mt Garnet, ANIC; Greenhills, ANIC; Home Hill, QDPIM; Ingham, ANIC, SAMA; Mt Inkerman, ANIC; Iron Rng, QDPIM; Julatten, ANIC; Julia Creek, SAMA; Kenilworth SF, UQIC; Kingaroy, ANIC; Kuranda, ANIC, SAMA; Laura, QDPIM; 70 km N Laura, QDPIM; 25 km N Laura, QDPIM; Mt Lewis, ANIC; Mackay, ANIC; Magnetic Isl, SAMA; Malanda, SAMA; 23 km E Marecba, QDPIM; Mary Creek, ANIC; Melvor, ANIC; 21 mis S Miriam Vale, ANIC; 100 mis S Miriam Vale, UQIC; Mission Beach, ANIC; 3 mis W Mourilyan, ANIC; Mornington Isl mission, SAMA; 5 km W Mossman, QM; Mossman George Nt Pk., UQIC; Mossman George, ANIC; Normanton, QDPIM; Oxley Ck, QM; Patuma Dam, ANIC; 9 km W Paluma, QDPIM; Prince of Wales Isl, NMV; 8 mis NE Proserpine, ANIC; Rocklea, QM; Stanthorpe, QM; Mt Spec, ANIC; Stewart Rng, SAMA; Tolga, SAMA; Townsville, ANIC, NMV, SAMA; 9 km ENE Mt Tozer, ANIC; Weipa, QDPIM; 144 C. H. S. WATTS Windsor Tableland, QDP1M; Yamba, UQIC; Yungaburra, QDPIM. South Australia Arkaba Creek, SAMA; Mambray Creek, SAMA; Marne River., SAMA; Mt Remarkable, SAMA; 6 km E Nilpinna, SAMA. Victoria Dimboola, ANIC; Genoa, ANIC; Sea Lake, SAMA. have, however, been unable to reliably separate light coloured females. In general E. deserticola are larger and more weakly punctured but considerable character overlap occurs between the two species. In more southern areas E. deserticola overlaps with E. eyrensis, E. samae and E. elongatus but can be separated from all three by the much weaker punctation and in the case of E. elongatus by the well-marked dark patterning on the head which E. elongatus lacks. Western Australia 6 km NNW Broome, ANIC; Beverley Springs, WAM; Cadjeput Rockhole, WAM; Charnley River, 25 miles N Beverley Springs, WAM; Derby, SAMA; Drysdale River, ANIC; 7 miles NE Giles, WAM; Gill Pinnacle, WAM; Hammesley Rng, SAMA; Koolan Island, WAM; 6.5 km N Mt Bell, WAM; Kununurra, ANIC, QDPIM; Millstream, ANIC; 5 km SE Millstream HS., ANIC: Mitchell Plateau, ANIC; West Peewah River, ANIC; Synnott Range, WAM; 14°53'S, 125°45'E, SAMA. Remarks A relatively large species showing considerable variation in size, colour and strength of dorsal punctation making its identification difficult without dissection. The two colour morphs are well marked and occur together in the same populations. Occasional specimens have an intermediate dark-testaceous dorsal surface to a variable extent. It is my impression that the dark form is more common in the north where it is the dominant form in many areas. The black morphs are readily identified by their weak to moderate punctation contrasting markedly with the strong punctation in the two other species with black morphs, E. eyrensis and E. samae. It is also the only species with the combination of tips of the parameres hooked, weak to moderate dorsal punctures and black patterning on the head. In the North-east the species overlaps with E. weiri and E. pseudoweiri. These can be separated from most E. deserticola by the lack of black morphs, somewhat stronger punctation, pale tips to the maxillary palpi and the small amount of black on the front of the head. An occasional specimen of E. deserticola may also have pale palpi tips and much reduced black on the front of the head. Over most of its range E. deserticola is sympatric with E. maculiceps. The two species are readily separated by the male genitalia and presence of black morphs in E. deserticola. I Enochrus (Methydrus) elongatus (W. MacLeay) Philhydrus elongatus W. MacLeay, 1 873 Enochrus (Lumetus) elongatus (W. MacLeay) Knisch, 1924 = Philhydrus caledonicus Fauvel, 1883: syn. nov. = Enochrus caledonicus (Fauvel): Knisch, 1924 Types Enochrus elongatus: Lectotype: 'Philhydrus elongatus Gayndah McL W'; 'K19506'. Right- hand specimen of two mounted on same card, AM herein designated. Paralectotypes: 1 , same data as lectotype, left- hand specimen, AM; 3, on same card, same data as lectotype, ANIC; 1, 'Philhydrus elongatus McL', 'Griffith Collection Id by A. M. Lea', 'Co- Type', SAMA; 1, 'Philhydrus elongatus M. Queensland cotype 14629', SAMA. Enochrus caledonicus: Lectotype: 'm', 'Nouvelle Caledonie. Noumea Marais Anse Vata, Octobre Saves, ex. coll. Fauvel', 'Coll. et det. A. Fauvel Philhydrus caledonicus FvP, '17— 479% 'lectotype' on red label. Remounted with genitalia dissected out, IRSNB. Paralectotypes: 1 , 'Nouvelle Caledonie Kanala, Rec. Coste, ex coll. Fauvel'; 'Coll. et det A. Fauvel Philhydrus caledonicus Fvl.' '17-479', 1, 'Nouvelle Caledonie. Anse Vata, Noumea Saves', 'Coll. et det A. Fauvel Philhydrus caledonicus Fvl,' IRSNB. Synonymy after examination of types. Description (number examined, 330) Fig. 7 Elongate oval, body flattish, length 4.4 mm - 5.0 mm. Dorsal surface testaceous, diffusely blotched, lighter towards sides, rear of head often diffusely darker but never distinctly darker than rest of head. Ventral surface dark red-brown, apical portions of legs lighter, palpi light testaceous, tips of maxillary palpi dark brown. Head covered with small sharply impressed, well AUSTRALIAN ENOCIIRUS 145 8 10 11 12 13 14 FIGURES l— 14. Dorsal views of apical portion of male genitalia. 1 & 2, E. maculiceps showing range of variation; 3, E. pseudoweiri; 4, E. deserticola; 5, E. eyrensis; 6, E. isabellae; 7, E. elongatus, lateral view of median lobe; 8, E. weiri; 9, E. samae: 10, E. euhenangeei; 11, E. aliciae; 12, E. malabarensis; 13, E. esuriens; 14, E. mastersi. 146 C. H. S. WATTS separated punctures, quarter to half size of eye facet, a dozen or so much larger punctures inwards from each eye. Pronotum similarly punctate, a row of systematic punctures about 3x diameter of adjacent punctures running inwards from each front corner and smaller group of similar punctures a little behind middle on each side. Elytral punctures slightly stronger than those on pronotum, stronger laterally than on disc. Serial punctures, larger, sparse, hard to trace towards apex. Underside evenly setose- punctate, rugose portions of femur, approximately three quarters length of femur. Mesosternal keel well developed, anterior edge curved somewhat, nearly perpendicular, ventral edge flat or slightly convex. Notch on apical stemite small but well marked. Male: Aedeagus with crotchet-shaped hook at tip. Tips of parameres outwardly hooked. Metaclaws with enlarged basal lobe and rest of claw straightened. Pro- and mesoclaws curved with thickened basal portion Female: Claws curved with thickened basal portions. Distribution Australian Capital Territory Black Mt, ANIC. New South Wales Albury, ANIC; Allyn River, NMV; Armadale, ANIC; Balranald, ANIC; Bathurst, AM; Bogan River, AM; 20 mis SSW Bourke, SAMA; Broken Hill, SAMA; Coonabarabran, AM, ANIC; Cowra, AM; Ecceleston, AM; Hay, ANIC; 37 km E Hay, SAMA; Mt Kaputar Nat. Pk., ANIC; Kosciusko Nt Pk., ANIC; Lake Menindie, SAMA; Menindie, SAMA; Narrabri, ANIC; Norfolk Island, ANIC; 97 km S Tibooburra, NMV. Northern Territory 53 km NE Alice Springs, ANIC; 12 km SW by W Alice Springs, ANIC; Top of Ayers Rock, WAM; 97 km N Barrow Ck, SAMA; 30 km NE by E Borroloola, ANIC; Ellery Creek, NMV; Finke, SAMA; Glen Helen, NMV; Stanley Chasm, ANIC; 15 km East, Vaughan Spring, HS, ANIC. Queensland Bollon Dist, AM; 42 km Boulia, ANIC: Bunya Mts, UQIC; Camooweal, QDPIM; Edgbaston HS, SAMA; 5 km NE Edungalba, SAMA; Glen Alpin, UQIC; Helidon, UQIC; 6 km E, Kamma, ANIC; Mackay, ANIC; MacPherson Rng, Nt Pk., AM; Marmor, ANIC; Mt Isa, QDPIM; Mt Spec, ANIC; Normanton, QDPIM; Tolga, QDPIM; Toowoomba, UQIC. South Australia Aroona Dam, SAMA; Mt Ban, ANIC; Old Billa Kallma HS, SAMA; Blanche Cup Spring, AM; Broken Hill, AM; Coopers Creek, SAMA; Coward Sp., SAMA; 16 km E Curdimurka, SAMA; 36 km ESE Curdimurka, ANIC; 1 1 km NE by Etadunna HS, ANIC; Gawler Rngs, SAMA; Lake George ANIC; Mt Lofty Rng, AM; Mannum, SAMA Marne River, SAMA, McDovall Peak, UQIC Murray Bridge, AM, SAMA; Olary, ANIC Oodnadatta, SAMA; Paralana Hot Springs, SAMA; Quorn, AM; Renmark, SAMA; Roonka Stn, SAMA; Roseworthy, SAMA; Scorpion Springs CP, SAMA; River Torrens, SAMA; Warburton, River, ANIC; Wearing Gorge, SAMA; Willochra Ck, SAMA. Tasmania 15 mN Waratah, UQIC. Victoria Eppalock Res., NMV; Eustace Gap Ck, NMV; King Lake Nt Pk., NMV; Kiata, NMV; La Trobe River, NMV; Lake Hattah, ANIC; Lilydale, SAMA; Little Desert, ANIC; Lake Hattah, NMV; Marysville, SAMA; Melbourne, UQIC; Preston, NMV; Shepparton, ANIC; Skeleton Ck, SAMA; Stawell, SAMA; Swan Hill, NMV; Traralgon Ck, NMV; Violet Town, ANIC; East Warburton, NMV; Wyperfeld Nt Pk, ANIC. Western Australia 15 km SSE Armadale, ANIC; 102 km SE by E, Broome, ANIC; 12 km NE Broome, ANIC; Bunbury, ANIC; 23 km NW by W Mt Arid, ANIC; 7 miles NE Giles, WAM; Gill Pinnacle, WAM; Great Victoria Desert, WAM; Mary sp. HS, ANIC; 13 km ESE Mooka H.S., WAM; Koolan Island, WAM; Serpentine River, NMV; West Peawah River, ANIC; Wilga, ANIC; Wmgello, ANIC. Remarks Enochrus elongatus differs from all other Australian Enochrus by the pale coloured dorsal surface including the head. It is the only Australian species with the tip of the aedeagus distinctly hooked upwards and with a basal swelling on the meta- and mesoclaws as well as the proclaws. The rear of the head is variably coloured. In many specimens it is darker than the front of the AUSTRALIAN ENOCHRUS 47 head, but diffusely so and never black and sharply delineated from the front as in most other species. Enochrus elongatus superficially resembles the pale morphs of E. deserticola and E. eyrensis in size and colour. It also shares with them, and no other Australian species, the hooked parameres. Unlike them the black colour morph common in these two species does not appear to occur in E. elongatus. In both E. deserticola and E. eyrensis the head is predominantly black with only small patches of yellow-brown on the front angles. In E. elongatus the head is uniformly yellow-brown or the rear is diffusely darker. In E. deserticola the dorsal punctation is much weaker than in E. elongatus and in E. eyrensis it is stronger. Enochrus elongatus is a widespread and common species within Australia including Norfolk Island. I can find no significant difference between Australian specimens and the types of E. caledonicus thus extending the range of E. elongatus to New Caledonia. Enochrus (Methydrus) esuriens (Walker) Philhydrus esuriens Walker, 1858 Enochrus (Lumetus) esuriens (Walker): Knisch 1924 Enochrus (Methydrus) esuriens (Walker): d'Orchymont 1927 = Philhydrus pullus Fauvel, 1883: syn. nov. = Enochrus (Lumetus) pullus (Fauvel): Knisch 1924 Types Enochrus esuriens: Type locality: Ceylon (Sri Lanka). Type not seen. Enochrus pullus: Lectotype: 6 'Coll. IRSNB, Nouvelle Caledonie, Anse Vata, Marais Juill Octobre Rec. Saves, ex. coll. FauveF; 'Coll. et det A. Fauvel Philhydrus pullus Fvl. IRSNB 17-497', IRSNB with red lectotype label, herein designated. Seen. Paralectotypes: 3 6 , 1 9 7 same data as lectotype in IRSNB. I have not seen type material of E. esuriens and have relied on the work of Balfour-Browne (1939, 1945) and specimens identified by him in BM(NH) for my concept of this widespread species which is known from India, through South-East Asia to Northern Australia and Polynesia. Description (number examined, 130) Figs 13, 18 Length 1.6 mm- 2.3 mm. Oblong-oval. Elytron and pronotum light-testaceous, often blotched and streaked darker, head black, area in front of eye, to about width of eye light-testaceous, central black panel two to three times width of lateral light area. Ventral surface dark-testaceous, appendages lighter towards extremities, whole of apical segment of maxillary palpi usually darker. Punctures on head moderately strong, systematic punctures inwards from eyes about size of eye facet and about twice diameter of adjacent punctures. Punctures on pronotum somewhat weaker, punctures on elytra somewhat stronger than on head, those towards apex and laterally about half diameter of adjacent serial punctures which are often not easy to trace. Mesosternal keel weakly developed. Male: Aedeagus narrow, pointed, virtually lacking terminal dorsal pad, shorter than parameres, collar nearer base than tip, parameres broad in basal half narrow for most of apical half, inner edge only very weakly sinuate, tips rounded. Proclaw straightened, strongly swollen at base, meso- and metaclaws rounded, swollen at base. Distribution New South Wales Wootton, ANIC. Northern Territory Adelaide River, ANIC; E Alligator River, AM; Darwin, ANIC; Humpty Doo, QDPIM; 12 km N Humpty Doo, QDPIM; 6 km N Humpty Doo, QDPIM; Jim Jim Creek, ANIC; 9 km N by E Mudginbarry HS, ANIC; 19 km E by S Mt Borradaile, ANIC; 12 km NNW Mt Cahill, ANIC; Nourlangie Creek, ANIC; Oenpelli, AM. Queensland Annan Falls, ANIC; Ayr, ANIC; 29 km S Bamaga, ANIC; Cairns, ANIC, SAMA; Calliope River, ANIC; Cardstone, ANIC; 60 km S Coen, SAMA; 40 km N Coen, SAMA; Ingham, ANIC; Lansdowne St, ANIC; 73 km NW Laura, ANIC; Laura, QDPIM; Macleod River, SAMA; Mackay, ANIC; Mareeba, QDPIM; Mary Creek, ANIC; Mt Baird, ANIC; Mt Coolum, ANIC; Mt Ivor River, ANIC; Mt Molloy, ANIC; 2 km NNE Mt Tozer, ANIC; 1 1 km ENE Mt Tozer, ANIC; 8 km E by N of Mt Tozer, ANIC; 12 km S Normanton, SAMA; North Pine River, UQIC; Port Douglas, ANIC; Redcliff, UQIC; 15 km WNW South Johnstone, QDPIM; Townsville, ANIC, SAMA; Weipa, QDPIM. 148 C. H. S. WATTS Western Australia Mitchell Plateau, AN1C; Nullagine, WAM. Remarks Enochrus esuriens' small size (<2.5 mm), predominantly black head and weak mesosternal keel make it one of the most readily recognised species of Australian Enochrus. It is most easily confused, not with other Enochrus, but with species of Paracymus and Anacaena, These latter genera lack the systematic punctures present on the head, pronotum and elytra of Enochrus. Enochrus (Methydrus) eubenangeei sp. nov. Types Holotype: