Journal of the Royal Society of Western Australia, Vol. 68. Part 1. 1985, p. 1-8. Destruction of australites by aborigines in part of tbe Eastern Goldfields, Western Australia by W. H. Cleverly and Evelyn I. Cleverly W.A. School of Mines and 79 Ward Street. Kalgoorlie. W.A. 643U Manuscript received 21 ■\ugust 1984: accepted 19 February 1985 Abstract Searches at 238 natural sources of drinkable water in part of the Eastern Goldfields of Western Australia have resulted in the discovery of australites at 13% of the sites, usually as flakes and in very small numbers. Already existing collections of australites from the study area totalling 26 609 specimens have been examined and the abundance of flakes determined. From these two studies and from a consideration of the forms of the flaked australites, it is estimated that less than 1% of australites have been used destructively by Aborigines. This level of destruction can have had no significant effect upon the australite distribution pattern. Introduction. It is conceivable that the use of australites (Australian tektites) by Aborigines (Baker 1957, Johnson 1963-4, Edwards 1966. Akerman 1975) could have affected significantly their numbers and distribution pattern, thereby rendering unreliable features of the distribution which, it is hoped, provide pointers to australite origins (Cleverly 1976). This study concerns the degree of destructive use within an area of about 70 000km^, or less than 2% of the visible sirewnficld. avoiding the question of removal of australites from their sites of find for non-destructive use as charms or ritual objects or tor trade. The study area (Fig. I) includes parts of the Menzies, Edjudina, Kalgoorlie, Kurnalpi, Boorabbin and Widgiemooltha 1:250 000 map sheets, SH51-5, -6, - 9. -10. -13 and -14 in the R102 and National Topographic map series. Geological maps with the same names on the same scale issued by the Bureau of Mineral Resources, Geology and Geophysics and by the Geological Survey of Western Australia provide some synonymous and additional place names. Where necessary, as for example with duplicated names, localities are referred to by map number and metric co- ordinates in the following style: — SH51-9, UF4687. The extent of australite destruction by Aborigines has been investigated by searches for flaked australites and by the examination of already existing collections for their presence. Results 1. Rock and australite flakes at natural sources of water Rock flakes foreign to the site are often found i^^ar gnamma holes and other natural sources of drinkab c water in the Eastern Goldfields, The flakes are usually varieties of opal or cryplocry'sialline silica such as common opal, moss opal, chalcedony or jasper which develop especially abundantly in this semi-arid climate as weathering products of ullrabasic rocks. Other materials such as cherts, fine grained quartzites, aphanitic igneous rocks and the silicified cappings oi sediments may also be represented. All of these are tough and their more or less perfect conchoidal fracture makes them generally suitable for working. The combination of suitability of materials, occurrence distant from their outcrops and close association with sources of water is strong circumstantial evidence that the materials were transported to the sites by Aborigines and are the debris or the unused fraction from the making of artifacts. The existence of aborigines in an extensive semi-arid region must have depended greatly upon their ability to find and use wisely the natural sources of water. It appeared therefore that natural sources of water, even if not recognized occupation sites, would at least have been visited from time to time and would be likely places at which to seek evidence of the use of australites. The results of searching water sources for introduced rock matter and australites are presented in Table 1 for the 32 sites where australite specimens were found. A concise statement of the same items of information for all 238 sites can be supplied on request. These sites are not an exhaustive sample of the natural water sources of the area but they do include well over 90% of the named granite rocks, soaks, gnamma holes, rock holes, pools and swamps. Rock flakes were not usually found immediately alongside the water but some tens of metres distant on a sandy rather than rocky slope, however slight, overlooking the winter whenever such a slope was available. The occurrences of introduced rock flakes have been roughly quantified in four categories of increasing abundance according to the quantity found per person in ten minutes, the search period commencing when the area of concentration, if any, had been located. The categories are:— rare one to three flakes uncommon more than three flakes, less than a handful common one to two handfuls abundant more than two handfuls Around some water points there was no evident favoured area, but an occasional flake, core stone, grinder or anvil stone (usually broken) was found thinly scattered over an area of up to a hectare or so. In such cases, the abundance was usually ‘‘rare” though the total amount of material might be considerable. 42580-1 1 Journal of the Royal Society of Western Australia, Vol. 68, Part I, 1985. Figure 1. Portion of the Eastern Goldfields of Western Australia covering parts of six 1: 250 000 map sheets (broken line boundaries): the name town of ^cn sheet is shown, rilled circles indicate water sources examined except that for those less than 4 km apart, only the more important is shown C rossed circles are sites where australite specimens were found. Sites of australite abundance are numbered thus:— 1. McAuliffe Well 2 Wangine Soak. 3. Carr Bovd Rocks. ^ Introduced rock flakes in widely differing abundance and degree of dispersal were found at 85 % of the searched sites. Of the remainder, two-thirds are within the extensive area of granite occupying the south-west quadrant of the study area, where introduced flakes are rare or absent except at marginal sites or where related to the large inlier of stratiform rocks centred on Ryans Find. The granite is so extensive that introduced materials would need to be carried at least 50 km to a site such as Thursday Rock. The absence of introduced rock flakes does not mean that the site was not visited by Aborigines, who may have used materials immediately at hand. For example, at a granite “rock" (island hill), they may have used quartz, aphanitic apophyses of the granite or deeply weathered and subsequently silicified granite crusts (a widespread material with some resemblance to quartzite in breaking through the enclosed quartz grains and siliceous matrix): there are 2 Journal of the Royal Society of Western Australia, Vol. 68. Part 1. 1985. conchoidally fractured pieces of such materials, especially the siliceous crusts, at some sites. The writers claim no ability to recognize artifacts but only to recognize rocks foreign to their surroundings. Thus a water source such as Gnarlbine Rock, a granite with excellent water seepages and a native well, was certainly used, though only two introduced rock flakes were found in two searches. Likewise, introduced rock flakes were not found at the Taurus gnamma hole (Table 1) which is within uhramafic rocks, but the superabundance of locally derived opal and chalcedony flakes might well include artifacts. Similarly again for a plunge pool at Smithficld where the surroundings are silicified sediments admirably suitable for use, and for a gnamma hole south of Jaurdi. where there is abundant outcropping jasper bar only 100 m away: there w'ould be no need to introduce raw materials from a distance to such sites. Table 1 Number of australites and abundance of rock flakes al natural water sources in the Eastern Goldfields Water source, map sheet, co-ordinates .\usiralites Registered number of australites: notes 'UF A 3rF L'larring Rock, an upstanding granite with extensive seepage areas, soaks (including rock-lined Government soak), gnamma holes and shallow holes. SH5 1-5. TG6387 0 2 u SM 1 2 096. Flakes thinly widespread. Clavpan. Shorty Dam adjoins. SH5 1-6. UH975 1 0 1 u SMI 1769. Flakes especially round south-west margin. Rock holes in Davis Creek. SH51-6, VH0843 0 7 u SMll 776. SM12 005. Flakes on low sandy rise overlooking creek. Prospector Pool in Nine Mile Creek. SH5 1-6, UH8 1 5 1 0 1 u SM 12 087. Flakes mainly on north-west side. Four Mile Pool. SH51-6, UH8344 3 5 u SM 1 2 088. Flakes on west side, rare elsewhere. Top Pool in tributary of Yerilla Creek. SH5 1-6, UH8744 0 1 u SM12 086. Flakes on bare areas south side of creek. McAuliffe Well. Rock-lined soak at foot of granite rock with also shallow holes on the rock. Rock arrangement nearby. SH51-6, IIH9637 10 200 A SM 1 1 704. Flakes plentiful on rock and al northern fool below the soak towards the rock arrangement. Also 4 australites and 1 15 flakes in Tilloison colls. Gnamma hole in low outcrops of granite. SH5 1-6, UG7388 0 1 u SM12 052. Gnamma hole and a few shallow holes in low granite outcrops at east end of Cockatoo Rocks. SH5 1-6. UG8685 0 R SM12 095. Very small pools in creek west of Princess Bore. SH51-6. VG0189 0 C SM12 106. Flakes common in small patches both sides of creek. W'albrook Swamp Dam, A large dam occupying most of a hollow, original nature now indeterminate. SH5 1-6, VG3 1 89. . 0 u SM 12 108. Flakes along northern shoreline. Wanginc Soak. Seepage with soaks and deeper wells at foot of breakaways. Also small plunge pools and other rock holes in the gullies incised in the breakaway edge. SH51-9, TG9362 6 204 A SMI 1 755. Rock and australiie flakes occur as lag deposit on low sand dunes overlooking the soak area from the south-east. Plunge pools (3) and smaller rock holes down the gully from the breakaway edge about 250 m wcsl-souih-wcst from Wanginc Soak. Also several shallow holes on the plateau surface. SH5 1 - 9. TG9362 0 2 U SMI 0 933. Rock flakes uncommon on the plateau in general vicinity and up to 200 m north- westerly along the former telegraph line to Davyhursl. Small granite rock in elbow of creek and wash therefrom northward towards Lake Gwen. SH51-9. LIGl 1 60 4 A SMI 2 029. Granite forming middle north shore of Lake Owen. SH51-9. UGI367 0 2 C SM12 027. Cane grass swamp about half kilometre long at narrow south- western end of very large lignum swamp in crab hole country and contiguous with it. Latter not searched. SH51-9. UG5556. 2 0 C SMI 2 105. Flakes common on north-west margin. Rare to uncommon on south side. Gnamma hole in duricrusi alongside Broad Arrow to Ora Banda road. SH51-9. IJG3133 0 U SM9532. 42580-2 3 Journal of the Royal Society of Western Australia, Vol. 68, Part 1. 1985. Table 1 — continued. Water source, map sheet, co-ordinates Australites Registered number of australites; notes 'UF “F 3rf Two small swamps with very sparse cane grass. SH51-9, UG4809 5 0 R SM878 i , SM 1 1 208. So-callcd “Little Gidgi". Granite rock with several gnamma holes, some shallow holes, wall system and dam (Eight Mile Rock Dam). SH51-9, UF0991 0 1 U SM 1 2 099. Flakes on sand at east side. Cane grass swamp. SH51-9. IJF51 87 0 U SMI 2 097. Flakes on gentle sandy slope at east end. Carr Bovd Rocks, an extensive granite with marginal seepages, shallow holes. SH51-10. IIG7776 23 A SMI 2 051. Flakes especially near south end but also thinly over more than 1 km northerly. Cane Grass Water Hole, aclavpan. SH5I-10. UG9570 0 6 A SMI 2 100. Flakes abundant round most of shoreline and across the floor. Wangalli Rock, a low granite with soak and several gnamma holes. SH51-10. VG147] 0 U SMI 1 777. Lake Emu, an extensive lignum swamp with also reeds, tea tree, cane grass in vicinity of the dam within the lake near its south end. SH51-10. UG8764 0 2 C SMI2 089. Hakes along more than 1 km of the .south-east margin overlooking the dam in deepest part of the lake. Binti Binti Rocks. Seepages at foot of breakaways. SH51-10. UG9757 0 1 U SMll 778. Stone arrangement on the plateau surface. Rock flakes in vicinity of seepages. Yowic Rock Hole, a gnamma hole in duricrust over granite near its low broken edge. SH5 1-10, VG.J729 0 1 C SMI2 098. Flakes on gentle slopes above the broken edge. Small lake 3.5 km west of Harper Lagoon. SH5 1-10. UG5920 0 2 u SMI 2 069. Hakes on north shore. Gnamma hole in duricrust near Taurus mining centre within ultra-basic bell with chalcedonic and opaline crusts. SH51-10. UG8702 0 1 — SMI 2 050. Introduced rock flakes not found but superabundant flakes of chalcedonic and opaline materials present, Gnamma hole in duricrust 2.5 km east of north of Golden Ridge. SH5I-10. UF7189 0 1 u SM9 1 34. On the old Kurnalpi coach road. Swamp with sparse cane grass and marginal tea tree. SH51-10, UF6879 1 0 c SMI 2 093. Flakes especially at south-west corner and western shoreline. Cowarna Rocks. Granite with two gnamma holes (larger formerly covered), marginal well, some shallow holes. SH51- 10, VF4077 1 5 c SM12 104. Rock flakes widespread, abundant in patches, especially down slope from gnamma holes. Karramindie Soak. Marginal to low granite outcrops shown on earlier maps as Fourteen Mile Rocks. Modern maps show Fourteen Mile Rocks 2 km to the south. SH51-I 3, UF4466 0 2 c SM9020, SMll 765. Rather contaminated by bottle glass. 'Unflaked australitcs, either whole or natural fragments. •Australite flakes and flaked cores. ■’Abundanceofiniroduced rock flakes; A = abundant, C - common. U - uncommon, R - rare. Certain searched sites have been omitted from the lists and are not therefore included in the statistics. The Bullock Holes (SH51-10, UG892I), an important source of water to gold prospectors, and the “Waterhole” near Feysville (SH5I-10, UF6375) are representative of those omitted because they were probably not significant water sources until deepening or other improvements had been carried out by the white man, whose artifacts (but not rock flakes) are found nearby. Further features which were cerlinly sources of water to Aborigines, even abundant sources, have sometimes been omitted because they were found to be so modified by quarry ing (e.g. Cardunia Rocks), extensively contaminated by bottle glass (e.g. Wallaroo Rocks) or generally contaminated that their assessment is difficult and the result unreliable. However, it has been possible to include a few features such as the gnamma hole on the Mungari granite (since destroyed by quarrying), the gnamma hole on the Ora Banda road (a regular tourist bus stop) and the contaminated Karramindie Soak because sufficient obser\'ations had been made on them as earlv as 30 years ago. At only 32 (13%) of the 238 sites were australites also tound, mostly as flakes and in very small numbers (Table 1 ). Even where plentiful, as at Wangine soak, they lorm an insignificant fraction of the numbers and mass of the rock flakes. At no locality were they found without rock flakes, at least of local derivation if not introduced. 4 Journal of the Royal Society of Western Australia. Vol. 68, Part 1. 1985. They were not found, except marginally, in the extensive area of granite and sand plain in the south-west quadrant and at sites within the Widgiemooltha map sheet. Australite flakes were abundant only at Wangine Soak and McAuIiffe Well, localities previously known (Cleverly 1976: 221): Carr Boyd Rocks is in a lesser category. .Australite specimens recovered during the present work have been registered in the collection of the Geology Department. W.A. School of Mines (SM). 2. Flaked aiistralites already present in collections The upper limit of destruction of auslralites by Aborigines may be estimated from representative located samples in collections if it is assumed that all flaked specimens are artifacts except those with brilliant vitreous lustre which are the likely results of “testing” (Baker 1957: 14) or some other recent cause of fracture. The Tillotson collections, which contain 9 946 specimens from the study area representing 50 carefully searched localities from near Widgiemooltha in the south to Lake Raeside in the north, are the only acceptable sample available for the area as a whole. The flake abundance and maximum possible artifact abundance is: — 277/9 946 or 2.8% Samples available for rather following results: — large areas gave the Hampton Hill Station 241/21 927 or 1 . 1 % Edjudina Station 67/1 874 or 3.6% Estimates for samples representing smaller areas are as follows: — Mount Remarkable Station .. 10/320 3.1% Boyce Creek, Yerilla Station 3/143 2.1% Run-ins to Black Flag Lake... 2/177 1.1% Seven Mile Hill Salt lake on Ora Banda pipe 1/299 0.3% track Kambalda and adjoining 0/300 — Lake Lefrov 0/195 — For the Mount Remarkable sample above, 8 of the 10 flakes were found around water sources. In wet seasons, Boyce Creek contains three large fresh-water pools. The Seven Mile Hill area includes the Afghan Rocks group of gnamma holes. No australite flakes were found in the two areas of salt lakes, which suggests that those from other areas are indeed the work of .Aborigines. Note on flaked auslralites Flaked australite specimens seen at water sources are similar to those already present in collections. It is unusual to find an australite with a single flake scar: the type has been illustrated by an example from outside the study area (Fig. 2B). Flake scars on opposed sides of an elongated form resulting in a chisel-like shape are also rare (Fig. 2C) but arc evidently widespread. There is a specimen of this type in the Finke, N. T. collection (SAM) and one from South Australia has been illustrated by Edwards U966 PI. 2 G,H). On the wider australites. two or more scars may be present on one or both sides (Fig. 2 D-J). Flakes are by far the most common form with some flaked cores. About 10% arc “cap pieces”, one surface being the curved and weathered outer surface ol the australite, the other the conchoidal fracture scar of detachment (Fig. 2 L). A further 60% show at least some small area of weathered outer australite surface. This often lakes the form of a narrow strip between two sub- parallel fracture sufaces (see upper end in Fig. 2 N), indicating that at least two flakes were removed from the australite. The remaining 30% have no remnant of outer surface. Some of the smaller specimens of this type show the scars of considerable work (Fig. 2 O, P, Q, Sf Consideration of the foregoing leads to a conclusion pertinent to this investigation. The iw'o commonest groups constituting nearly 90% of specimens show' that at least two and often several flakes had to be removed to account for their shapes. Thus any one flake is not likely to represent a distinct individual australite, though it could be true for occasional “cap pieces”. Several flakes could represent the work or even part of the work done on a single australite. Discussion Searches of 238 water sources resulted in the discovery of 511 australite specimens of which 31 are whole or naturally fractured and 480 are flakes or flaked cores. .As distinct from the rock flakes, which are by definition imports to the sites, the auslralites might either have been found nearby or brought in. The high proportion (more than 1 5:1) of flaked to unflaked specimens (sec for example Fig. 2A). their occurrence at water sources and invariable association with rock flakes suggest that most, if not all flaked auslralites were shaped by aborigines. Experience in other parts of the australite-strcw-n field supports this view. Edw'ards (1966) classified 443 specimens from Aboriginal camp-sites in South Australia into 130 complete specimens (29%). 161 fractured specimens without purposive trimming (36%), 56 trimmed pieces (13%) and 96 implements (22%). Concerning the group of untrimmed fractured specimens. Edwards said “Most of these arc flakes with a well developed bulb of percussion. Since they w'cre collected on former camp-sites they probably w'cre produced by human agency”. Thus three groups totalling 71% could be attributed to Aboriginal workmanship and the remaining 29% complete specimens were “believed to have been used in Aboriginal ‘magic* ”. Akerman (1975) classified 385 pieces from around a gnamma hole near Rawlinna into 295 struck flakes. 60 utilised flakes and 30 implements, thus regarding all specimens as artifacts. Akerman ( 1975) also examined 137 specimens (SMIO 943-5) from Spider Bore and the adjoining Mesquite Swamp on Earaheedy Station, finding 72 struck flakes. 5 used flakes, 6 flaked cores and 34 implements, a total of 85% of specimens being thus regarded as artifacts. The 480 flaked specimens found during the present work include 427 from three sites where there were evidently special reasons for their popularity. A localized tribal custom seems unlikely because a defined tribal boundary for which there is good evidence (Tindale 1974: 143, 252 and map) separates Wangine Soak and Carr Boyd Rocks south of the line in Maduwongga country from McAulilTe Well, just north of the line in Ngurlu country. The ready availability of large auslralites may have encouraged their use at McAuliffc Well. The 167 auslralites from the general vicinity available in collections include 7 in the high weight range 20.8-42.5 g. The average weight of complete specimens is 6.23 g and of all specimens 4.72 g, more than twice the averages of 2.99 g and 1.96 g respectively for the 26 609 Eastern Goldfields specimens examined. 5 Journal of the Royal Society of Western Australia. Vol. 68, Part 1. 1985. tt* •.y^ •«. I I igurc 2. Flaked australiics Irom Western Australia, natural si^c except m item A. A. Three essentiallv complete atislraUlcs at upper kTl largest mm I II Hakes or llaked cores, the product of 10 person-hours collcciihg at McAuiiffe Well. Yenlla Station SMTi 704 r %;‘'|*'‘'Tal}le with single flake scan Spider Bore. Earaheedy Station. SMIO 944. C. Two views of an elongated auslralilc llaked ni one end Viu'"? t"- Two v;.ews of flaked australUe. Luke Emu. G.ndalb.e Station. SMI2 089. E. Two v^-w.of flaked ausKalUe S ? rn It australjje Eaurus cnlla Station SMI 2 088. W I wo views ol pointed fragment. Mc.AulitTc Well. SM 1 I 704. As pointed out by Edwards (1966), there are two principal drawbacks to the use of australites — their small size and the inferior propenics of the glass compared with cry ptocrystalline forms of silica. The first of those disadvantages may not have applied with the usual force at McAuliffe Well. Setting aside the Hakes from the three sites of abundant usage, the remaining 53 were found in small numbers at 25 different sites. There are at least two possible reasons why australites were not found at water sources in the extensive area of granite and sand plain in the south-west quadrant of the study area. The simplest is that australites did not fall there: the shower is generally admitted to have been very “patchy" and there do not appear to be any australites from the area in collections. However, it would be a considerable coincidence if a distribution feature of the 6 Journal of the Royal Society of Western Australia, Vol. 68, Part 1. 1985. shower should coincide, even approximately, with a geological boundary'. It is therefore more likely that auslraliles fell in the area and have been buried in the extensive eluvium derived from the granite or in the drifting sands. A perusal of the School of Mines records shows that several australitcs from other parts of the sand plain were indeed found in post-holes, pipe trenches or borrow pits. Australitcs have been found marginally, for example at Karramindie Soak (Table 1 ). There are australitcs in collections from various places within the Widgiemooltha sheet but there are few evident water sources. The failure to recover flaked australitcs from within that part of the study area might therefore be related to inadequacy of the sample. The priman- concern in this study is with destructive usage but brief comment is made here upon the 31 complete and naturally fractured australitcs (those having fracture surfaces as weathered as the primary and secondao' surfaces). They constitute 6% of recovered specimens, about one-fifth of their abundance amongst a comparable number of specimens from South Australian camp-sites (Edwards 1966). Only 6 of the 31 weigh more than 4 g. have a maximum dimension 20-25 mm and could probably have been utilized for the production of small implements. That they were not so used suggests the possibility of retention for some non-destructive role. The remaining specimens average 1.85 g and have correspondingly smaller dimensions. Some of them, especially those found in situations such as fresh water claypans, might well have been seen by .Aborigines and Ignored or discarded as too small, narrow' or mis-shapen for use. No evidence was recognized of any specimen having been collected for ritual or “magic" unless it is the obser\'aiion that the 10 unflaked specimens from McAuliffe Well include a rare “square-ended" aberrant form (Cleverly 1982). a naturally fractured specimen of the same type and a rather large and well preserved teardrop. It might be speculated that because of their unusual shapes these specimens were spared for some non-destructive — possibly ritual — purpose: there is a stone arrangement at the site. The first section of this paper concerning australite search may now be summed up. The forms of the specimens as discussed in the preceding section of this paper suggest that the 480 flaked specimens could have been derived from only one or two hundred australitcs. an insignificant number from an area which has yielded many thousands to dealers and lapidaries and more than 33 000 to known collections. It is a very' small yield when seen as the result of searching 238 likely places. Consider next the already existing collections. Flaked australitcs comprise 0-3.6% of various localized samples, not all of which arc mutually exclusive. The mean flake abundance is given by: — 466/26 609 or 1.75% This is the upper limit for abundance of australite artifacts. The only earlier attempt to assess quantitatively the destruction of australitcs by Aborigines is that of Baker (1957: 13) who noted “the extreme rarity (less than 0.5 per cent) of worked australite fragments amongst the large number so far recovered". Baker’s statement elsewhere (1957: 8) that Aboriginal chipped flakes and implements constituted “something in the order of 0.005 per cent" of the 30 000 to 35 000 australitcs in collections is clearly erroneous because 30 australite artifacts constituting about 0.1% of specimens were discussed in his paper. The erroneous statement may be reconciled with the acceptable one by omitting from it the words “per cent". The 30 000-35 000 specimens should not be confused with the more than 33 000 from the study area, at least 80% of which are in private collections which were unknown and/or unavailable to Baker. The basis of Baker's estimate differs from that used here. Strict criteria w'cre applied for recognition as an artifact — "... one can only be reasonably sure that certain fragmented australitcs were w'orked by .Aboriginal man if undisputed evidence is present of the application of marginal pressure flaking or the like" (Baker 1957: 13). It seems probable that little more than that fraction which other authors class as “implements" was accepted as artifacts. In contrast, no criteria whatever have been applied here, the upper limit of artifact abundance being thus determined. There is no available information on the number of implements compared with total flakes in collections c.\cept in the highly biased samples from occupation sites. There could be a considerable difference in the proportions because flakes found in occupation sites are likely to have been produced by human agency or at least attributed to it whilst those from elsewhere have more chance of being products of temperature changes (including grass fires), breakage by animals or vehicles, development of saw-cuts or some other weathering process. It is safe to say that Baker’s "less than 0.5 per cent" needs to be increased several times, possibly to "less than 3 per cent” to be placed on the same basis as the maximum of 1.75% found here. Conclusion Flaked australitcs from part of the Eastern Goldfields average 1 .75% of the australitcs in localized samples and could have been produced from the destruction of distinctly less than 1% of their number. The fact that collectors sometimes ignore flakes (Cleverly 1976: 220) is not applicable to the major collections used here. Samples collected from water sources do not suggest any need to increase the estimate. Even if all the flaked specimens arc artifacts, the level of destruction w'ould be too low to affect the numbers and distribution pattern as currently known. This estimate for part of the Eastern Goldfields is of the same order of size as the estimate of Baker (1957) for the australite-strewn field as a whole. The low level of usage suggests a possibility not initially visualized nor investigated that australitcs passing through the hands of dealers and lapidaries with virtually no written record, and the considerable number of poorly documented and therefore almost valueless specimens in collections (Cleverly 1976: 222) might well represent a greater loss of information on the australite- sirewnfield than is attributable to destruction by Aborgines. Acknin\U'iJ^cnu'nh. — W'c ihank ihc following persons lor ready access to pastoral propenies and/or assistance in reaching water sources: — Messrs B. M. Agars and B A. Agars (Pinjin). Mr and Mrs P ( artcr {Mount V'ellcrs). Mr J. F. Cotter (Bmnonngie). Mr C Cox {Rivenna), Mr and Mrs C Day (Yindi), Mr and Mrs 1. Duncan (Mcnanginal. Mr and Mrs P R. Egerton- Warhurlon (Mount Burges) Mr 1 1 Eldneh. Mr E H. Finlavson (.leedamya), Mi and Mrs W (icnsch (C.arbine), Mr and Mrs W. Gorry (Cow-arna Downs), Mrs I. h. Halford (Credo), Mr and Mrs L. J Johnson (Goongarne). Mr and Mrs B. C. Jones and B h. C . Jones (Hampton Hill), Mr R. A. C. Jones (Mount Monger), Ms C Lacey. Mr T. N. l.owc (Mount Remaikahic). Mr and Mrs L. J. McKav (Morapoi). Messrs B. McKay and K. McKay (Edjudina). Mrs ,1 I- Piercey (Walling Rock), Mr B. D. F. Robinson (Yen)la). Ms S. Ryan, Mr and Mi's John Fonkin and Mr and Mrs Stephen Tonkin (Gindalbio), Mr M L Wearnc. W'e thank also tlie following curators or owners of australite collections for loans of specimens: — Dr W' O Birch (National Museum of Victoria), Dr R- Hutchison (Brilnsh Museum (Natural History)). Mr and Mrs B, C. Jones. MrC B. C. Jones. Mr J. L. C. Jones. Mr L D. TiUoison. Mi and Mrs R. (l. 1 lIlolSOTl, Mr M. K. Quarlermainc took part in some field work and also processed our photographs used in Figure 2. Ms J. M. Wearne drafted Figure 1. 42580-3 7 Journal of the Royal Society of Western Australia, Vol. 68, Part I, 1985. References Akerman. K. (1975). — The use of australiies for the production of implements in the western desert of Western Australia. Univ. of Queensland Anthorpological Museum Occasional Papers No. 4. Baker, G. (1957). — The role of australiies in Aborginal customs. Mem. Nat. Mus. Via.. No. 22: 1-26. Cleverly. W. H. (1976). — Some aspects of auslralite distribution pattern in Western Australia. Rec. West. Ausl. Mus.. 4: 217-239. Cleverly, W H. (1982). — Some aberrant australiie forms from Western Australia. J. Roy. Snc. West. .4usf.. 65: 1 7-24. Edwards. R. (1966). — Australiies used for aboriginal implements in South Australia. Rec. S. Ausl. Mus.. 15: 243-250. Johnson, J. E. (1963-4). — Observations on some Aboriginal campsites South Australia and adjoining stales. Mankind. 6(2 & 4); 64-79 154-181. Tindale, N. B. (1974). — Aboriginal tribes of Australia. Univ. California Press. 8 Journal of the Royal Society of Western Australia, Vol. 68, Part 1. 1985, p. 9-12. The effect of wave action on the shell morphology of Littorina imifasciata Gray by G. Basingthwaighle and W. Foulds Western Australian College of Advanced Education. Claremont Campus W.A. 6010. Manuscript received 17 November 1981: accepted 18 March 1985 Abstract The aperture length/spire height ratio was determined for populations of Littorina imifasciata up a vertical slope in an exposed and a projected site. It was found that a significantly larger ratio occurred in the lower levels of the exposed habitat than on all the levels in the sheltered site. There was no difference between levels at the two sites. It is argued that the change in mean phenotype of this continuously variable trait may be the result of natural selection produced by varying degrees of wave action and desiccation. Introduction The common gastropod mollusc Lottorina imifasciata Gray is found on the rocky shores of Western Australia as far north as North-West Cape (Wilson and Gillett 1 979) ranging from the supra-littoral down to the upper tidal zone. Of all the grazing molluscs present on the vertical intertidal rock walls of Rottnesl Island, it is found uppermost in the vertical range (Black et al. 1979). Living in the intertidal zone subjects the fauna to a variety of environmental factors which result in gastropods displaying a wide range of morphological adaptations. Major environmental factors thought to be responsible for shell variations in gastropods are wave action, prolonged submersion, high temperatures, extreme salinity, desiccation (Slruhsaker 1968 Newkirk and Doyle 1975) and predation by animals such as crabs (Hughes and Elner 1979). The shell shape in the Western European Dog-Whelk, Nucellus lapuilus has long been known to vary with exposure, animals with short squat shells being found on exposed headlands whilst those w'ilh elevated and sharply pointed spikes are restricted to sheltered inlets (Kitching el al. 1966, Berry' and Crothers 1968). The shell-shape ratio (length of aperature: height of spire) has been shown to be directly related to an exposure scale devised by Ballantinc (1961) in southern and Western parts of Europe (Crothers, 1974, 1975a, 1975b, 1977, 1981: Crothers and Cowell 1979). However, populations in certain parts arc quite different (Crothers 1981), and even though in all these places the shells are more elongated than would have been expected from the regression it cannot be assumed that the correlation is universal in this particular species. The present preliminary investigation was intended to establish whether shell-shape ratio in populations of L. imifasciata is in any way related to the degree of wave action (exposure) on vertical rocky shores. Materials and methods The two study sites from which the populations were measured were on exposed and a protected section of rocky shore at Point Peron, Western Australia. The different degree to which the two rock faces are subjected to wave action were determined by counting the approximate splash height reached by each of 150 waves during a high tide period.* The two sites are opposite sides of a vertical limestone w'all; the exposed side faced the open sea while the protected side was situated in a protected cove. Samples were taken from each site at the low water mark, which was the lowest level the snail was found, the upper limit of the splash zone and an area in the middle of the two limits. The range from lower to upper was approximately one metre. Such a small scale was possible because patterns of tidal influence occur on a scale of centimetres on veitical shores characteristic of coastal limestone of Western .Australia. Also littorincs shift their position little after grazing excursions relative to tidal conditions. (Black et al. 1979). All three levels of both sites were sampled at the same lime on three occasions. 25 April, 1 1 May and 4 June 1981. During the first two collections 70 shells were measured from each of the SIX areas while 140 were collected on the last date. The maximum height of the spire (H) and length of the aperture (L) was measured to the nearest 0. 1 mm for each snail using vernier calipers (Fig. 1). The results were expressed as the length/height ratio in order to eliminate the variation in size due to differing ages of the individual specimens. L. unifasciata has a planktonic larval phase. The juveniles settle on the lower shore levels after metamorphosis and migrate upwards as they grow. The ratio of aperture length/spire height vanes with size. Therefore, to eliminate differences in the size frequency characteristics between exposed and protected shores being caused by differential settlement or size selective mortality only individuals with a spire height above 7 mm were measured. *Thi.s meihod of measuring wave action docs not measure ihe force exerted by ihc waves (a simple technique is not available) nor is it an objcclive meihod. A more quantitive technique was allcmpled u.sing Calcium sulphate clods (Doty D7I) but these were quickly destroyed on the exposed site. There was also no way of devising an exposure scale such as that of Ballantine (1961) as little floral or faunal zonation existed along the vertical slope. 9 Journal of the Royal Society of Western Australia, Vol. 68, Part 1, 1985. Figure I. — Aperture length (L)/spirc height (H) measurement of L. umfasciata. The six populations were tested for significant difference using the two-tailed “Z" statistic. It should be noted that any differences observed were due to shape and not size as the aperture length and spire height values were highly correlated for both exposed and protected sites (r=0.87 and 0.90 respectively). Results The results of Table 1 indicate that the degree of splash by wave action differs between the exposed and protected sites and along the vertical gradient. Table 1 Average number of times the zones are covered or splashed (min‘‘) at high tide. Site Upper level Middle level Lower level Mean Protected O.I 0.5 5.0 1.9 Exposed 1.0 5.3 8.6 5.0 Each determination is the mean of ! 50 measurements At the exposed site the whole population of L. unifasciata has a significantly larger ratio than the population sampled at the protected site (p<0.001). It can also be seen that at both sites there is no significant difference in ratio between any of the three vertical levels (Tables 2 and 3). The snails situated at the sheltered site have a slightly smaller but significantly different ratio than the snails from two lower positions at the exposed site (p<0.05). Table 2 Aperture lenglh/spire height ratio of 1. umfasciata populations collected from three vertical levels at a protected and an exposed site al Point Peron. Site Upper level Middle level Lower level Mean (A) IB) (C) (G) Protected 0.655 ± 0.657 4- 0.650 ^ 0.654 4 0.003 0.003 0.003 0.002 Exposed (D) 0.670 ± (E) 0.679 ± (H) 0.676 + 0.003 0.003 0.003 0.002 Table 3 “Z” Scores and levels of significance of L. unifasciata population comparisons on the three levels at the two sites. The operator error factor was 0.2 mm for aperture length and 0. 1 mm for spire height calculated to Xy ± 0.008 Figure 2 shows that in all six populations the ratio had a continuous distribution and most likely it is polygenic. The medians of the sheltered situations were skewed towards the lower ratios compared with those from the exposed site. Discussion As with many intertidal gastropods throughout the world (Stephenson and Stephenson 1954, Berry and Crothers 1968. James 1968, Kilching et al 'l968, Struhsaker 1968, Vermcij 1973, Newkirk and Doyle 1975) there appears to be a change in shell shape correlated with habitat in the Australian snail, L. unifasciaia. The selection pressures acting on the parameters measured seem to be different in the two extreme habitats, and greater between sites than between levels. On the exposed rocks where there is a great deal of wave action, there must be selection for morphological features which decrease the turbulence of water flowing over the shell. There must also be selection for increased area of contact with the substrate. This can be achieved by increasing the size of the aperture relative to the height of the spire. Such a morphological change is evident when one compares the exposed populations at Point Peron. subject to heavy wave action, with the more sheltered populations. The populations on the lower levels at the exposed site, particularly arc subject to more intensive wave action. Although on protected shores there is not so much splashing as on exposed shores. Black ei al (1979) clearly showed a gradient in desiccation stress on vertical rocky shores of Roitnest Island. The result is selection to reduce the size of the aperture relative to the height of the shell. However, there is no significant difference in shell shape between the levels at the sheltered site. Thus although desiccation may be a selection pressure it does not seem to be as important as wave action but it appears to eliminate the possibility of predators such as crabs being a method of selection as these are purported to be more abundant on protected regions (Crothers 1968, 1970). Irrespective of the relative importance of the selection pressures, it is not obvious whether the environmental influence is developmental orgenctical. However, variation of all populations shows a continuous distribution and the changes in shell morphology are by a directional shifting of the median left or right (Fig. 2) suggesting a genetical contribution. 10 frequency Journal of the Royal Society of Western Australia, Vol. 68, Part 1 , 1985. 100 50 EXPOSED A S HELTE RED D 1 00 B E 60 -65 70 -75 80 ratio F Figure 2. — Frequency of aperture length/spire height ratio classes of L. unifasciata. 11 Journal of the Royal Society of Western Australia, Vol. 68, Part 1, 1985. Acknowledgetnenis. — We thank Mr. P. McMillan for assistance. Mrs. S. de la Huniy for help with statistical analyses and Dr. M. Johnson for discussions of this study. References Ballantine. W. J. (1961). A biologically defined exposure scale for the comparative description oi rocky shores. Fiela Studies 1 (3); 1-19. Berry. R. J. and Crolhcrs, J. H. (1968). Stabilizing selection in the dog whelk (Nucella laptUus). J. Zooi. London. 155: 5-17. Black, R.. Fisher. K... Hill. A., and McShanc, P. (1979). Physical and biological conditions on a steep intertidal gradient at Rottnesl Island. Wesiem Australia. Aust. J. Ecoi. 4: 67-74. Crothers. J. H. (1968). — ^The biology of the shore crab Cardnus maenas{L): 2 The life of the adult crab. Field Studies. 2; 579-614. Crothers, J. H. (1970). — The distribution of crabs on rocky shores around the Dale Peninsula. Field Studies. 3: 263-274. Crothers, J. H. (1974). — On variation in Nucella lapdlus {L.y. Shell shape in populations from the Bristol channel. Proceedings of the Malacological Society of London. 41: 157-170. Crothers, J. H. (1975a). — On variation in Nucella fapillus (L). Shell shape in populations from the south coast of England. Proceedings of the Malacologtcal Society of London, 41: 489-498. Crothers, J. H. (1975b). — On Ihc variation in Nucella lapillus (L): Shell shape in populations from the Channel Islands and northwest France. Proceedings of the \falacological Society of London. 41: 499-502. Crothers. J. H. (1977). — On variation in Nucella lapillus (1.)'. Shell shape in populations towards the southern limit of its European range. J. Moll. Stud.. 43: 181-188. Crolhcrs. J. H. (1981). — On variation in Nucella lapillus (L): Shell shape in populations from Orkney and the north coast of Scotland. J. Moll. Stud..41: 182-189. Crothers. J. H. and Cowell. E. B. (1979). — On variation m Nucella lapillus (L): Shell shape in populations from Fensfjorden. J Moll. Stud.. 45 : 108-114. Doty. M. S. (1971). — Measurements of water movement in reference to benthic algal growth. Botanica Marina. 14 : 32-35. Hughes. R. N- and Finer. K. W. (1979). — Tactics of a predator Cardnus maenas. and morphological responses of the prey, Nucellus lapillus. J. Anim Ecoi. 48 : 65-78. James. B. L. (1968). — The characters and distribution of the subspecies and varieties of Littorina saxaiilis (Olivi, 1792) in Britain. Can. Biol. Mar.. 9: 143-165. Kitching. J- -A.. Muntz, L.. and Ebling, F. J. (1966). — The ecology of Lough Inc. XV. The ecological significance of shell and body form in Nucella. J. Anim. Ecoi. 35: 1 13-126. Newkirk. Cl. F. and Doyle. R W’. (J975). — Genetic analysis of shell shape variation in Liltonna saxaiilis ou an Environmental Cline. Marine Biology. 30; 227-237. Stephenson. T. A., and Stephenson. A. (1954). — Life between the tide- marks in North America. III. Nova Scotia and Prince Edward Island; description of the region. / Ecoi. 42: 14-45. Slruhsaker. J. W. (1968). — Selection mechanisms associated with inira- spccific shell variation in Littorina picta. Evolution. 22; 459-480. Vermeij, G. J. (J973). — West Indian molluscan communities in the rocky intertidal zoncr a morphological approach Bull. mar. Sd.. 23; 351- 386. Wilson, B. R. and Gillen, K. ( 1 97 1 ). — Australian shells. Read Pub. 12 Journal of the Royal Society of Western Australia, Vol. 68, Part 1, 1985, p. 13-15. Ecology of the large indigenous earthworm Megascolex imparicystis in relation to agriculture near Lancelin, Western Australia by I. Abbott, *J. S. Ross and C. A. Parker Soil Science and Plant Nutrition Group, University of Western Australia, Nedlands, W.A. 600'-) •Present address; Institute of Forest Research. Hayman Road. Como. W.A. 6 1 52 Afanuscripl rccctml /V f-c’hruarv 19S5: accepted !5 April Abstract The largest species of earthworm known in Western Australia was studied in agricultural soils near Lancelin. In pasture its frequency of occurrence in quadrats, density and biomass (fresh weight) were 18-20%, 5.0-5. 5 m'^ and 41-67 gm'“ respectively. Although indigenous, we failed to find it in soil under nearby native heath ('‘sandplain") or in soil from which native vegetation had recently been removed. Despite virgin and recently cleared soils differing from pasture soils principally in extractable P and N, eanhwoims cultivated in virgin and pasture soils in the laboratory showed no differences in survival or weight over 50 days. Ploughing of pasture did not cause an immediate reduction in density, but one year later (and after a second ploughing) density averaged 1 . 1 m*f Introduction Much of the agriculturally-developed land of Western Australia is occupied by two species of earthworm of European origin ( Abbott and Parker 1980). However, in 1980 Dr. D. L. Chatel drew our attention to an abundant population of a large earthworm (up to 30cm in length) captured by the tynes of a cultivator on a farm adjacent to Karakin Lake, 100 km NW of Perth. According to Mr. J. Wood, owner of “Karakin" farm, these large earthworms were present when native vegetation was cleared for agriculture in 1958. This earthworm proved to be the indigenous species Megascolex imparicystis Michaelsen 1907, which occurs between Dongara, Dandaragan and Perth (Abbott 1982). We aimed to determine the density and biomass of this species in virgin and agriculturally-deyeloped soils, and to explain why a population of a native species of earthworm could persist in agricultural soil. A previous survey (Abbott and Parker 1980) had shown that no native species occurred in agricultural soils of the wheatbelt of Western Australia, although several natjve species have been recorded in soils under native vegetation at Jilakin Rock and near Wongan Hills and Hvden (Abbott, Parker and Milewski. pers. obs.). Study area and methods In 1981 and 1982 we sampled quantitatively earthworm populations on pasture, recently ploughed pasture, virgin heathland and recently cleared heathland. Soils from each area were collected and characterized for selected physical and chemical properties. Environment The study area lies on the Karrakatta landform unit (Churchward and McArthur 1980), which consists of deep sands overlying aeolianite. In places limestone is exposed at the soil surface. These soils arc infertile because of their great age and the strong leaching action of heavy winter rains on their porous sandy surfaces (Bettenay et al. I960). Although P is essential for the establishment and maintenance of pastures, it is not the only factor limiting growth, as Cu and Zn deficiencies are'important especially after superphosphate has been added (Bettenay e/ a/. 1960). The climate of the study area is typically Mediterranean, with hoi. dry summers and cool, wet winters. At nearby Lancelin. the average annual rainfall is 627 mm, with nearly half falling in June and July. The whole area was virgin heath and woodland until 1958. Vegetation present on the virgin study site is heath up to 4 m tali. Principal species are Banksia atfenuata, R. sphaerocarpa. Dryandra sessilis. Uakea sp., CahthamniiS sp.. Eucalyptus todtiana, Xanihorrhoea preissii. Nuytsia Jlorihunda, and Allocasuarina humilis. The recently cleared site has a few individuals of Macrozamia riedlei. N. florihunda, X. preissii and Eucalyptus decipiens. It was cleared by chaining in July 1979. and burning in March 1980, planted with an oat crop in June 1980. grazed by sheep and then planted with another oat crop in 1981. The pasture site has a few clumps of Eucalyptus todtiana, E. gomphocephala and E. decipiens remaining, but is sown to Serradella [Ornithopus saliva) and Subterranean Clover {Trifoltum subterraneum). In June 1981 it had been last ploughed 6- 7 years previously. Since 1958. the pasture has received superphosphate and trace elements including Mn. Zn, Co and Cu. The recently ploughed site was sampled six days after ploughing in June 1981. It was then seeded with oats with a combine seeder, superphosphate being added also. In 1982 it was reseeded with lupins. All sites arc within I km of each other. Eield observations Fifty soil samples (19 x 19 x 29 cm depth) were randomly extracted by spade from each of the four sites in June ! 98 1 and 1 982. This soil was sorted by hand for earthworms, which were taken to the laboratory for weighing. Earthworms were also collected from tynes of the cultivator in order to determine mean weight of a larger sample. 13 Journal of the Royal Society of Western Australia, Vol. 68. Part 1 , 1 985. Soil analyses Soils were analysed for % gravel, coarse sand, fine sand, silt and clay (pipette method. Piper 1947), soil moisture (gravimetrically determined). pH (5:1 w/v watensoil), organic carbon (Walkley-Black, see Piper 1947), Total N (Kjeldahl method, see Piper 1947), extractable N (Purvis and Leo 1961), Total P (Murphy and Riley 1962), extractable P (Watanabe and Olsen 1965) and total K (HCl digest, see Piper 1947). Laboratory experiment A completely random design of two treatments (pasture soil, virgin soil) with five replicates was followed. The soils were obtained from the surface 15 cm of the pasture and virgin heath sites, air dried, and passed through a 1.96 mm sieve. Five lots of 5 kg each of virgin soil and pasture soil were placed in plastic pots (diameter 20.5 cm) lined with plastic bags. The depth of soil was 25 cm. The water holding capacity of the soils was determined and the soils in each pot maintained at 60% water holding capacity (Piper 1947). Pots were left to Stand in the laboratory for four weeks before the addition of earthworms. Earthworms were collected from the top 20 cm of soil close to Karakin Homestead and were selected to have similar weight. One worm was added to each pot. Worms were weighed (.Abbott and Parker 1981) every ten days for 50 days. Results Fresh weight p/'Megascolex imparicystis In June 1981, 52 apparently undamaged earthworms were collected from the ploughed paddock. Mean weight (± SE) was 8.21 ± 0.40 g, with range of 2.74-16.0 g. The frequency distribution of weights was: weight (g) 2-5.9 6-9.9 10-13.9 14-16 N 10 26 15 1 % 19 50 29 2 Mean weight of the June 1982 sample, collected from more fertile soil adjacent to the Homestead, was 12.2 ± 1.4 g (N=19), range 6.7-18.4 g with one individual weighing 27.4 g. Mean weights of undamaged specimens collected from the soil samples in the pasture were 4.5 g (N=8)in 1981 and 6.3 g(N=6) in 1982. These, however, are probably undcr-esli males of the true means because larger individuals tend to be cut into fragments during the extraction of the soil core. The method of obtaining worms from lynes would tend to miss smaller individuals and bruise larger ones. Table 1 Frequency, density and biomass (fresh weight) of Megascolex imparicystis in the four sites studied Site Frequency (%) Density (m ^) Biomass (gm"^) Virgin 0(0) 0(0) NA Recently cleared 0(0) 0(0) NA Pasture 1 8(20) 5.0(5. 5) 41.1(67.1) Recently ploughed.. 16(4) 4.4(1. 1) 36.1(13.4) NA. not applicable In each column, the 1981 figure is presented first, with the 1982 figure in parentheses. Frequency, density and biomass In both years no earthworms were found in the virgin or recently cleared sites (Table 1). In 1981 the frequency of occurrence of earthworms in pasture was similar to that in the recently ploughed pasture, but the latter declined by June 1982. The density and biomass of earthworms in 1982 in the recently ploughed pasture was also less than in 1981 (Table I ). Physical and chemical characteristics of soil Of the features examined (Table 2), extractable N was greatest in pasture and extractable and total P were greatest in both pasture and recently ploughed pasture. Laboratory experiment In both treatments worms showed an increase in weight over the first ten days (Table 3). However at no time during the experiment was there any significant difference in mean weight between earthworms in virgin or pasture soil. Discussion The lack of significant differences in weight between earthworms cultivated in virgin and pasture soil in the laboratorv' indicated that soil differences in extractable P and N do not alTecl the growth of Megascolex imparicystis. But paradoxically, the field investigations showed that no earthworms were recorded in virgin soil or recently cleared soil. Without the laboratoo' experirnent the obvious conclusion was that this species of native earthworm benefited from increases in extractable N and P (and possible trace elements) in pasture soils. The difficulty of sampling around tree roots in virgin soils may have biassed the sampling. If earthworms fed close to roots, thev may not have been detected. However, this factor should' not have been important in the recently-cleared site. Because of the fanner's observation of {he occurrence of this species before farming began, we conclude that Megascolex imparicystis is very sparsely distributed in virgin soils. It is uncertain whether growth rales in pots would be the same as in natural soil, but it would be almost impossible at this stage to do growth experiments in the field because these earthworms arc known to burrow down to at least 4 m. In the virgin soil the food resources would be derived from the partial decomposition of leaf and root material from the native plant community. In the pasture soil, where the earthw-orms are quite numerous, it must be assumed that the organic matter derived from subterranean clover and sheep manure is the chief source of food. The biomass of Megascolex imparicystis in pasture ranged from 40-70gm•^ i.e. 400-700 kg ha’’. This is about 2-3 times that of the sheep on the same pasture (stocking rate averages 4 ha'', J. Wood, pers. comm.). This earthworm biomass is similar to that obtained in pastures elsewhere in Western Australia (McCredie 1982, 1. Abbott, unpubl.). 14 Journal of the Royal Society of Western Australia, Vol. 68, Part 1, 1985. Table 2 Physical and chemical characteristics of the soil at the four sites studied Feature Virgin Recently cleared Pasture Pasture recently , ploughed Soil moisture (%) 7.1 (-) 8.4 (_) 9.9 (-) — (— ) Gravel (%) 5.6 (— ) 7.3 (-) 9.3 (— ) 4.9 (— ) Coarse sand (%) 86.5 (— ) 79.6 (-) 84.9 (88.2) 88.6 (— ) Fine sand (%) 10.0 (— ) 14.7 (— ) 1.1 (6.9) 0.) (— ) Sill + clay (%) 3.5 (— ) 5.7 (— ) 14.0 (5.6) 11.3 (— ) pH 5.94 (6.25) 6.66 (6.31) 6.38 (5.90) 6.00 (5.83) Organic C (%) 0.98 (0.75) 1.39 (1.10) 0.86 (0.81) 0.90 (0.97) Total N (%) 0.054 (0.030) 0.056 (0.060) 0.056 (0.056) 0.060 (0.043) Extract. N (ugg"’) — (58) — (58) — (72) — (56) Total P(ugg'') 26 (20) 22 (35) 50 (79) 56 (55) Extract. P(ugg'') — (3.3 — (2.3) — (5.3) 0.12 (6.2) Total K (Me/ lOOg) 0.17 (0.13) 0.17 (0.15) 0.22 (0.15) (0.12) — not determined In each column, the 1 98 1 figure is presented first, with the 1 982 figure in parentheses 1 able 3 Mean weight in grams { ± 95% confidence interval) of Megascolex imparicystis cultured for 50 days in virgin and pasture soils from near Lancclin Soil type Time (days) 0 10 20 30 40 50 Virgin Pasture 1.43 (0.31) 1.44 (0.33) 2.12 (0.43) 2.07 (0.53) 2.27 (0.47) 2.05 (0-45) 2.25 (0.53) 2.13 (0.41) 2.41 (0.45) 2.00 (0.37) 2.27 (0.43) 2.03 (0.39) Of interest is the absence from this pasture of Aporrectodea (=Allolohophora) trapezoides and Microscolex duhius. Both species are widespread in agricultural areas of Western Australia (Abbott and Parder 1980); however, they occur at “Karakin’' only in the homestead garden and around the sheepyards. It is not known if these species are excluded from the pasture through competition with Megascolex imparicystis or have not had time to occupy this land. Acknowledgements. — We thank Mr. J. B. Wood of “Karakin” for allowing us io study and collect earthworms on his farm, T. McCredie for helping with the sampling in 1982, and L. Wong for the mechanical and chemical analyses of the soils. References .Abbott, 1. (1982). — The distribution of earthworms in the Perth metropolitan area. Records of the Western Australian Museum, 10: 11-34. .Abbott. 1. and Parker, C. A. (1980). — The occurrence of earthworms in the wheat-belt of Western Australia in relation to land use and rainfall. Australian Journal of Soil Research. 18: .343-352. Abbott. 1 and Parker. C. A. ( 1981). — Inleraciions between earthworms and their soil environment. Soil Riologv and Biochemistrv. 13: 191-197. Bettenay, E.. McArthur. W. M. and Hingston. F. .1. (I960). — The soil associations of pan of the Swan C'oaslal Plain. Western Australia. CSJRO Div. Soils, Soils and Land Use Senes. No. .35 (24 pp). Churchward. H. M. and McArthur. W, M. (1980). — Landforms and soils of the Darling System Western Australia. In Atlas of Natural Resource.'^ Dtirfing System U estern Australia pp. 25-33. Department of f'onscrvation and Environment. Western Au.stralia. McCredic, T. (1982). — The ecology of the earthworm Aporrectodea trapezoides (Duges, 1828) in a Western Australian pasture. Research Report. Department of Soil Science and Plant Nutrition. University of Western Australia. Murphy. J. and Riley. T. (1962). — A modified single solution method for the determination of phosphate in natural waters. Anal. Chim. Acta. 27: 31-36. Piper. C. S. (1947), — Soil and Plant Analysis. University of Adelaide, Adelaide. Purvi.s, E. R. and Leo. M. W. M. (1961). — Rapid procedure for estimating poieniially-available soil nitrogen under greenhouse conditions. Journal of Agricultural and hood Chemistry. 9: 1 5. Watapabc, F S. and Olsen, S. R. (1965). — Test of an ascorbic acid method for determining phosphorus in water and NaHC03 extracts from soil. Soil Science Society of America Proceedings. 29: 677-678. 15 Journal of the Royal Society of Western Australia, Vol. 68, Part 1, 1985, p. 17-20. A further find from the Youndegin meteorite shower by J. R. De Laeter* and D. J. Hosie School of Physics and Geosciences. Western Australian Institute of Technology, Bentley. Western Australia, 6102. *Honorary Associate. Western Australian Museum. Manuscript received 17 June 19H5: accepted 20 August 1985 .Abstract An iron meteorite weighing 4.665 kg has been part of a minerals collection at Quairading District High School for many years, where it is used as an integral resource in a teaching module on meteorites. The cobalt, nickel, gallium and germanium contents of this specimen have been determined by X-ray fluorescence spectrometry. It is identified as a member of chemical group lA (Wasson 1974). A detailed examination of the microstructure and chemical composition of this specimen with respect to other Youndegin meteorites confirms that it is part of the Youndegin meteorite shower. Introduction The first recorded meteorites in Western Australia were a number of irons discovered by Alfred Eaton towards the end of the 19th Century when agriculture was being established to the east of the early settlement at York. These meteorites became known as the “Youndegin" meteorites after a police outpost, located between the present locations of Cunderdin and Quairading. Four iron meteorites were found approximately 1.2 km north-west of Pikaring Rock (Figure I). These specimens weighed 11.7 kg. 10.9 kg, 7.9 kg and 2.72 kg and are now known as Youndegin 1 to IV respectively. The meteorites were found on the surface within a few metres of each other. Weathering products of the meteorites were also found in the immediate vicinity, suggesting that the specimens were part of a single shattered or disintegrated mass which had resided on the earth's surface for a considerable period of time. In 1891 a much larger specimen, 173.5 kg in weight, was discovered to the south-east of Pikaring Rock, and in 1892 yet another large specimen weighing 927 kg was discovered. These two meteorites were named Youndegin V and VI respectively. However some meteorites found in the same district were not given the name “Youndegin”. In 1 892 two iron meteorites were found to the east of Pikaring Rock. These weighed 92.3 kg and 0.68 kg and were given the name “Mount Stirling”. Other meteoritic fragments named Mooranoppin were subsequently found to the north of Pikaring Rock, although the exact location is uncertain. The largest meteorite in the district was found in 1903 in the Wamenusking area south-east of Quairading. Its existence was not officially known until 1952 when Mrs W. Sharett sent a picture of the meteorite to the “West Australian” newspaper. It was subsequently donated to the Western .Australian Museum by Mr E. C'. Johnson in 1954. This 2 626 kg meteorite was given the name “Quairading”. Youndegin Vll, a 4.1 kg iron meteorite, was found in 1929 approximately 8 km north-east of Pikaring Rock. Other fragments from this area, weighing a total of 13.6 kg, have been found from time to time in the same vicinity, and are collectively known as Youndegin VIII. Simpson (1938) suggested that the Mooranoppin and Mount Stirling meteorites were pan of the Youndegin meteorite shower. He also described Youndegin VII and VIII. and pointed out that one of the pieces of Youndegin VIII w'as made into a horseshoe by a blacksmith in York. McCall and De Laeter (1965) provided details of all the meteorites listed above. They pointed out that Quairading was in all probablity pan of the Youndegin meteorite shower. McCall (1972) reports that there is a small sample of iron shale, collected prior to the removal of the Quairading meteorite to the W..A. Museum, in the collection at the Western Australian School of Mines. De Laeter (1973) made a detailed examination of the geographical location, microstructure and chemical composition of the Mount Stirling, Mooranoppin and Quairading meteorites and compared them with samples of Youndegin l-VIL In particular the cobalt, nickel, gallium and germanium contents of these ten meteorites were determined by X-ray fluorescence spectrometry. The similarity in the chemical data enabled all ten meteorites to be classified as members of chemical Group LA (Wasson 1974). De Laeter (1973) concluded that the specimens were all part of the Youndegin meteorite shower which probably resulted from a meteoroid travelling in a south westerly direction. Table 1 gives details of the various meteorites found in the Quairading district, their masses, date of find and the present location of the main mass of each specimen. 17 Journal of the Royal Society of Western Australia, Vol. 68, Part 1, 1985. Table 1 Details of the Youndegin meteorites Name Main mass Date of find Location of main mass Youndegin I 11-7 kg 1884 British Museum. London: Youndegin H 10.9 kg 1884 9.82 kg National Museum, Youndegin III Melbourne: 10.9 kg 7.9 kg 1884 Western .“Xustralian Youndegin IV 2.72 kg 1884 Museum: 5 kg British Museum. London; Youndegin V 173.5 kg 1891 .2.7 kg Field Museum. Chicago; Youndegin VI 927 kg 1892 141 kg Naturhislorisches Youndegin VII Museum. Vienna: 927 kg 4.1 kg 1929 Western .Australian Youndegin VIII..,. 13.6 kg 1891-1929 Museum; 3.9 kg Private Collections Ward-Coonley Collection; 1-Ikg Western Australian Mooranoppin 1.6 kg 1893 0.820 kg Mourn Stirling Quairading 92.3 kg 0.680 kg 1892 Museum 0.725 kg Australian Museum. Sydney: 67.2 kg 2 626 kg 1903 Western Australian Museum: 2 626 kg 18 Description of the new find Quairading District High School (previously Quairading Junior High School), has had a number of meteorites donated to it by farmers and students in the district. Most of these specimens were small fragments about the size of a 20 cent piece, and none now remain in the School. However two large samples were retained . by the School’s Science Department. The meteoritic fragments were donated to the School prior to 1 972. Mr K. Ireland, a science teacher at Quairading District High School, has produced a leaching module on meteorites to capitalize on the unusual situation of a school being in possession of meteorite specimens, and located in an area in which many meteorites have been discovered. The module is part ot an astronomy topic, which is itself a sub-set of a science course taken by all i secondary school students in Western Australia. The School was prepared to allow the two specimens to be examined at the Western Australian Institute of Technology. The two specimens weighed 1.130 kg and 4.665 kg respectively. However when the specific gravity of the specimens were measured, values of 4.19g cm'^ Journal of the Royal Society of Western Australia, Vol. 68. Part 1, 1985. and 7.40g cm'^ were obtained respectively. It was therefore obvious that the smaller sample was not an iron meteorite. In all probability it is a piece of iron ore. Two photographs of the larger specimen are shown in Figure 2. The meteorite is approximately 16 cm long, by 14 cm wide, by 8 cm higli. A polished and etched surface of the new find is shown in Figure 3. The ^4CM Widmanstatten pattern may be compared to Figures 2 and 3 in Dc Laeler (1973). The etched surface is ver>' similar to these other samples of the Youndegin meteorites. The meteorite can therefore be classified as a coarse (Og) or coarsest (Ogg) octahedrite, (Buchwald 1975). The main constituent is tlie nickel-iron alloy kamacite, arranged in regular, well-defined plates, parallel to the faces of a regular octahedron. The apparent thickness of the plates varies from 1 to 5 mm. The specimen exhibits the richness of inclusions described by Simpson ( 1 938) and Buchwald ( 1 975). Figure 2. — The 4.665 kg iron nieleoriie from the Quairading District High School. Figure 3. — A polished and etched section of the new specimen of the Youndegin meteorite shower. 19 Journal of the Royal Society of Western Australia, Vol. 68. Part 1, 1985. Table 2 Analytical Data for the Youndegin meteorite shower Meteorite Nickel {%) Cobalt (%) Gallium (ppm) Germanium (ppm) Reference 7.08 + 0.04 0.45 ± 0.01 90 ± 3 342 + 8 This work This work De Laeter(1973) This work De Laeter (1973) This work De Ueter(I973) 6.79 ± 0.04 0.44 ± 0.01 84 ± 3 340 + 8 360 + 8 346 + 8 359 ± 8 348 ± 8 6.81 ± 0.04 6.85 ± 0.04 0.44 ± 0.01 0.45 ± 0.01 87 + 3 90 + 3 6.83 ± 0.04 6.81 ^ 0.04 0.44 ± 0.01 0.45 £ 0.01 90 ± 3 90+3 6.83 ' 0.04 0.46 ± 0.01 90 + 3 346 ± 8 The first detailed examination of the Youndegin meteorites was made by Fletcher (1887) who described the presence of schreibersite and a cubic form of graphitic carbon which he named cliflonite. Buchwald (1975) has described the microstructure of the Youndegin meteorites in detail. In addition to schreibersite. Buchwald points out that iroilite occurs as scattered inclusions, often intcrgrown with significant amounts of graphite. Cohenite is also common, and often forms rims around the schreibersite inclusions. Chemical composition Table 2 gives the nickel, cobalt, gallium and germanium concentrations determined by X-ray fluorescence spectrometry on flat, polished pieces of a number of specimens of the Youndegin meteorite shower, using the technique described by Thomas and De Laeter (1972). The previous determinations of the four elements by De Laeter (1973) are also listed in Table 2. The errors quoted with the values are based on counting statistics and arc at the 95% confidence level. The correlation between the values for cobalt, gallium and germanium for the new specimen and for Mount Stirling, Quairading and Youndegin III is extremely good. The nickel value of 7.08 ± 0.04% for the new specimen is higher than for the other three samples analysed in this study. De Laeter (1973) analysed 10 samples of the Youndegin meteorite shower and obtained a range of nickel values from 6.47% to 6.92%. The sampling problem with coarse octahedrites can be quite difficult. Wasson (1970) obtained a nickel value of 6.38% for a Youndegin sample, whilst a value of 7.4% was obtained for Mount Stirling (Wasson 1974). Buchwald (1975) suggested that the low value obtained by Wasson (1970) could have been measured on a specimen in which the kamacite was more abundant than for the bulk meteorite. Conversely, high nickel values may result from specimens with more abundant taenite than in the bulk meteorite. One of the disadvantages of X-ray fluorescence spectromeir> is that the measured concentrations arc only representative of a thin surface layer of the specimen. Wasson (1974) defines chemical group lA as those iron meteorites with 190 to 520 ppm germanium which fall within main sequence fields on germanium-gallium and germanium-nickel plots, with nickel and gallium values in the range 6.4% to 8.7% and 55 ppm to 100 ppm respectively. The analytical data listed in Table 2 confirm that the new specimen is a member of chemical group lA. Thus the chemical composition and microstructure of the new specimen from the Quairading District High School confirms that it is a fragment of the Youndegin meteorite shower. Youndegin is an impressive shower comprising numerous fragments ranging up to 2 626 kg in weight. The approximate extent of the shower as shown in Figure I covers an approximate area of 25 x 15 km (discounting Mooranoppin whose location is uncertain), with most of the samples being found in the , vicinity of Pikaring Rock. It is unfortunate that no firm details are available as to the location of this new i specimen of the Youndegin meteorite shower since it is the first new specimen to be reported for approximately 50 years. However one cannot rule out the possibility that other specimens are in existence, and perhaps some of these may be located in the future. Acknowledgements. — 1 would like lo thank Mr K. Ireland, a science teacher at Quairading District High School, for providing the meteorite described in this paper. Mrs P. R. Harris provided valuable technical assistance for the project, whilst Mr W. H. Cleverly commented on an earlier draft of the paper. References Buchwald, V. F. (1975). — Handbook of Iron Meteorites: Their History. Distribution. Composition and Structure. University of California Press, Berkeley. Dc Laeter, J. R. (1973). — The Youndegin meteorite shower. Meteoritics. 8: 1 69-1 79. Fletcher. L. ( 1 887). — On a meteoritic iron found in 1 884 in the Sub-district of Youndegin, Western Australia. Min. Mag.. 7: 121-130. McCall, G. J. H. and De Laeter. J. R. (1965). — Catalogue of Western Australian Meteorite Collections. Western Australian Museum, Perth. McCall. G. J. H. (1972). — Second Supplement lo "Catalogue of Western Australian Meteorite Collections '. Western .Australian Museum, Perth. Simpson, E. S. (1938). — Some new and little known meteorites found in Western Australia. A//>i. Mag.. 25: 157-171. Thomas. W. W. and De Laeter, J. R. (1972). — The analysis of nickel, gallium and germanium in iron meteorites by X-ray fluorescence speewomeiry. X~Ray Spectrometry. 1; 143-146. Wasson, J. T. (1970). — The chemical classification of iron meteorites — IV. Irons with Ge concentrations greater than 190 ppm and other meteorites associated with Group 1. Icarus. 12; 407-423. Wasson, J. T. (1974). — Meteorites. Springer Verlag. New York. 20 Journal of the Royal Society of Western Australia, Vol. 68, Part I, 1985, p. 21-25. Host distribution, potassium content, water relations and control of two co-occurring mistletoe species by Byron Lament School of Biology, Western Australian institute of Technology Bentley, 6102. Australia Manuscript received 20 Sovemher J03 m high were assessed for number and identity of mistletoes visible from the ground or > 15 cm wide, height (clinometer), diameter at ground level, and closest dstance of that tree from a tree bearing a mistletoe {Amyeuia miquelii if it was a eucalypt, Amyema pretssii if it was another species) until 80 trees were scored. Whole mistletoes were harvested from the major host species, returned to the laboratory in plastic bags and sorted into leaves, flowers or berries and major (including haustorium) and minor stems and their fresh weight taken. Fresh weights of samples were then obtained, dried at SOT for at least 48 h, milled and digested with cone. Potassium concentration was determined by flame photometer after addition of La. Cs and HCl (Lamonl and Southall 1982a). The process was repealed for leaves and minor stems obtained from the host tree. Total K content of the mistletoe was obtained from the concentration and dry weight figures for each organ. The water rcser\mir of each mistletoe was calculated in the same way. Percentage water content was obtained from twigs treated as for whole mistletoes above. Replicate twigs were used for determination of xylem pressure potential (portable pressure bomb) and osmotic potential (sap collected from pressure bomb by filter paper and placed in a sample chamber/microvoltmclcr). Other twigs were inserted into Griffin potometers for determination of transpiration rates over 30 min. Canopy temperatures were measured with an air thermister/telethermometer. 21 Journal of the Royal Society of Western Australia, Vol. 68, Part 1, 1985. Table 1. Incidence of mistletoe species on 80 trees in relation to potential host species, dimensions and distance from other trees bearing mistletoes. Data are x ± s, where appropriate. Nearest mistletoe (last column) refers to distance of each tree from the nearest tree bearing a mistletoe. Potential host species No. trees measured % with mistle- toes No. per tree Maximum No. per tree Tree height (ni) Tree diameter (cm) Nearest mistletoe (m) 1 . Presence oi Amvema preissii 5.5 ± 0.9 16.4 ± 6.1 Acacia podalyriaefolia 22 73 7.7 20 19.4 ± 26.5 Acacia hailevana 5 60 4.2 17 10.6 ± 4.6 32.0 ± 16.0 34.5 ± 42.9 Acacia mearnsii 5 80 1.6 3 8.8 ± 1.7 32.9 ± 21.7 10.7 ± 8.8 .■icacia pvenantha 2 100 I.O 1 8.6 ± 1.6 29.5 ± 8.1 10.3 ± 10.3 Acacia decurrens 1 100 1.0 1 8.3 43.0 10.0 2. Presence otAmvema miquelii Eucalyptus calophylla 24 54 1.8 14 15.6 ± 5.7 57.3 ± 28.0 36.6 ± 38.6 Eucalyptus marginata 8 1 0.1 1 11.7 ± 4.3 62.3 ± 36.0 13.7± 6.1 3. Absence of either mistletoe Dryandra scssilis 3 0 0 0 4.8 ± 3.2 10.1 ± 3.7 40.9 - 51.4 Eucalyptus camaldulensis 3 0 0 0 1 1.2 ± 2.9 43.5 ± 27.2 53.7 ± 40.1 Eucalyptus hotryoides ? 0 0 0 8.7 * 3.9 20.1 ± 15.4 70.0 ± 42.4 Eucalyptus maculaia 2 0 0 0 7.4 ± 0.6 17.7 ± 2.4 59.5 ± 57.3 Eucalyptus cladocalvx 1 0 0 0 13.2 22.5 56.0 Eucalyptus astringens Eucalyptus globulus 1 0 0 0 1 1.0 29.0 33.0 1 0 0 0 6.6 7.1 50.0 relative humidities by thermohydrograph and wind speed by hand-held or digital anemometers. Leaf areas were determined later with an electronic planimeter and doubled for isobilateral leaves or multiplied byTT for terete leaves. Results are means (x) ± standard deviations (s) throughout, even though the data did not always fit a normal curve. Seeds of Amyema preissii were squeezed from their berries and placed on the underside of branches 10-40 mm wide. Individual trees of 17 species received 100 seeds each. The number of mistletoe seedlings and leaves per mistletoe were assessed over 60 weeks. Results Mistletoe Distribution Table I shows that 51% of trees > 4 m high carried mistletoes at the study site, Amyema preissii was borne by 74% of Acacia trees and Amyema miqueUi by 35% of Eucalyptus trees. While some plants of all seven Acacia species, including A. longifolia and A. saligna not reported in Table 1. were parasitized, nine of 11 Eucalyptus species, including E. lehmamni. E. leucoxylon and E. rohusta not in Table I, possessed no mistletoes. Acacia podalyriaefolia was the most heavily infected species with on average 7.7 mature mistletoes per tree up to a maximum of 20. This was followed by Acacia bailevana (4.2) and Eucalyptus calophylla (1.8). E. calophylla was the most common tree ai the site (30%), followed by A. podalyriaefolia (28%), E. marginata (10%), A. baileyana and A. mearnsii (both 6 %). Apart from Dryandra sessilis, few plants of which exceeded 3 m, .1. podalyriaefolia was the shortest tree (5.5 ± 0.9 m) and had the narrowest trunk diameter (16.4 ± 6.1 cm). Among the eucalypts. E. calophylla was much larger (15.6 ± 5.7 m tall. 57.3 + 28.0 cm tnjnk diameter) than those species not bearing mistletoes. There was no (non-parametric) correlation between average number of mistletoes per individual and mean distance from the nearest neighbour bearing at least one mistletoe. However, there was a strong negative correlation between the percentage of individuals of each species carrying mistletoes and its mean distance from the nearest mistletoe-bearing neighbour (p < 0.00 1 , Spearman’s rank). The major exception was Eucalyptus 22 marginata where only one of eight trees possessed a mistletoe despite a mean distance of 13.7 ± 6.1 m from the nearest source of Amyema miquelii. Trees of E. calophylla carried on average over 14 times as many individuals of this mistletoe but were almost three times as far from the nearest source. Table 2 Variation in dimensions of Eucalyptus calophylla and Acacia podalyriaefolia trees with and without mistletoes. Data are x ± s. Significances refer to Mann-Whiiney tests. 1 . Eucalyptus calophylla Amyema miquelii Present Absent Sig. No. of trees 13 11 No. of mistletoes . .. 3.3 0.0 *** Tree height (m) 17.7 ± 6.1 13.1 ± 4.1 *** Tree diameter (cm). 66.8 ± 28.1 46.0 ± 24.4 *** Closest mistletoe (m) 29,4 ± 38.2 45.3 ± 39.0 *** 2. Acacia podalvnaefolta Amyema preissii Present .Absent Sig. No. of trees 16 6 No. of mistletoes .... 7.3 0.0 *** Tree height (m) 5.6 ± 0.8 5.2 ± 0.9 NS T ree diameter (cm). 17.1 ± 0.6 14.5 ± 0.7 *** Closest mistletoe (m) 19.4 ± 29.8 19.2 ± 16.8 NS Within a species, trees with mistletoes were much larger than those lacking them (p < 0.001, Table 2). This association probably did not show up for height of Acacia podalyriaefolia because most plants lacking mistletoes were <: 3 m, the minimum height for inclusion in the study. Trees of Eucalyptus calophylla with mistletoes were much closer to neighbours bearing mistletoes than those without (p < 0.001), but this was not true for A. podalyriaefolia. Potassium content The outermost branches, especially the leaves of both mistletoe species, contained much higher levels of K (Table 3) than equivalent parts of the five host species examined (p < 0.05 for minor stems and leaves, Wilcoxon matched-pairs test). There was a gradient Journal of the Royal Society of Western Australia, Vol. 68, Part I, 1985. Table 3 Potassium levels in leaves and stems and reproductive parts (when present) of two mistletoe species on five host species at Kalamunda Railway Reserve. Minor stems refer to the outermost branches of the canopy. Units are mg K/gdry tissue. Data are single values or x±S. Species Major stems Minor stems Leaves Flowers Berries Overall Eucalyptus calophylla. — 3.6 5.1 Ainvema miquelii 3.2 6.3 22.3 17.4 — 9.0 Eucalyptus marginata 2.75 2.7 Amvema miqudii 2.25 3.7 16.3 13.5 — 7.0 .Acacia podalvnaefoHa 3.7±2.4 4.0+1. 0 Amvema preissii 4.0±1.3 8.6±1.7 15.0 + 5.2 — 14.9 + 5.1 11.9±3.9 ■Acacia pycnantha 2.5 5.3 . Xmvema preissii 7.3 13.6 26.6 — — 18.7 Acacia decurrens 9.0 1.3 Amvema preissii 7.2 13.7 39.0 — 37.6 — increase of K from the haustorium (major stems) to the leaves, with a slight decrease in the flowers and fruits (including stalks). The overall concentration of K in Amycma prcLssii (15.3 mg g"') was almost twice that in Amyema miquelii. Whole mistletoes represented a K drain on the host of 4. 28 ±8.11 gfor.I. and 5.02 ± 2.87 g for A. miquelii. The largest mistletoe collected {A. preissii on Acacia pycnantha) contained 22.20 g of K. ^Vaier reiadon.s The outermost branches of both mistletoe species consistently contained about 10% higher levels of water (Table 4) than equivalent parts of six host species (p < 0.05. Wilcoxon matched-pairs test). Whole mistletoes contained 49.0 + 10.6% water, which represented a wmer volume of 1.77 ± 3.06 1 in a total weight of 4.19 ± 7.69 kg for both species. The largest mistletoe collected (Amyema preissii on Acacia pycnantha) weighed 19.78 kg and contained 7.9! 1 of water. Xylem pressure potential (^x) was consistently much lower in the mistletoes than in the six host species examined (p < 0.05. Wilcoxon). The osmotic potential of extracted sap (0o) was also lower in the mistletoes (p < 0.05, Wilcoxon). Table 4 Water content, xylem pressure and osmotic potentials of twigs of two mistletoe species and six host species. Data are x + s. Negative signs are omitted from potential values. Species Water content (%) iff X d*ar) •A 0 (bar) Eucalyptus calophylla 51.4 10.8 + 3.8 3.4+2.0 Ameyema miquelii.. 59.9 13.8±9.7 3. 7 + 2.3 Acacia podalyriaefolia .... 46.2±7.0 17.9 + 3.4 4.8±l.8 Amvema preissii 56.6±6.6 23.6 + 2.2 5.5±1.6 .Acacia pycnantha.... 59.7±9.8 9.9+2.7 2.6+ 1.8 Amyema preissii 72.5+14.4 17.5+12.0 4.4±1.4 .Acacia bailcvana..... 57.5+16.0 12.1+12.6 4.6 0.2 Amvema preissii 70.9±9.l 19.1 + 18.2 6.2+4.0 Acacia mearnsn 55.3±5.5 9. 1+3.8 5.4±2.4 Amyema preissii 62.3±3.4 13.6 + 2.3 4.4+0.3 Acacia decurrens 63.6 7.8 3.0 Amyema preissii 70.8 9.6 4.5 Table 5 Comparison of water relations between Eucalyptus calophylla with and without mistletoes and other Euca/yprus species lacking mistletoes. Results are ranges. Negative signs omitted from potential values. Sig. refers to Mann-Whiiney tests. T ree species Mistletoes Water content {%) tfi X (bar) ^o(bar) A. Eucalyptus calophylla present 51.4 5.5-14.6 2.2-6.4 B. Eucalyptus calophylla...,,, absent 50.3-58.4 5.2 2.3 C. 6 Eucalyptus Spp..., absent 38.3-66.6 7.2-21.4 1. 7-7.3 Sig. (A V Cj — «* NS There was no difference apparent in % water, j/f ^ and •P o between the trees of Eucalyptus calophylla with and without mistletoes (Table 5). However, there was a tendency for the other (non-host) eucalypt species to have lower ^ values than E. calophylla (p < 0.0! , Mann-Whitney test), but no differences in % water or *P O- Except for Eucalyptus marginata, the other six host species had much lower rales of transpiration (Table 6) than the two mistletoe species (p 0.05, Wilcoxon matched-pairs lest). Omitting Amyema miquelii on E. marginata. whole mistletoes lost 29.43 ± 30.07 ml water per hour from 10.00 to 15.00 h over the two days measurements were taken. Conditions were overcast, rain falling overnight though insufficient to moisten the dry surface soil. Host canopy lempcralure averaged 23X (range I4.2-2S,4‘'C), relative humidity 73% (range 58- 100%) and windspeed 1.2 m s'^ (0-2.5 m s'*). Host specificity Of the 1700 seed of Amyema preissii removed manually from their berries and placed on the various tree branches. 93.1 ± 5.7% germinated. By eight weeks, the adhesive pad of almost all germinants was fixed to the bark (Table 7). By 28 weeks, this had fallen markedly on most host species and by 60 weeks, no seedlings remained alive on all 10 Eucalyptus species and Dryandra sessilis. Some seedlings were still alive and growing vigorously on all five Acacia species while most seedlings were barely alive on Nuyisia jloribunda. Amyema preissii was most successful on Acacia podalyriaefolia, with 72% establishment and 58 ± 70 leaves after 60 weeks. 23 Journal of the Royal Society of Western Australia, Vol. 68, Part 1. 1985. Table 6. Transpiration rales of twigs of two mistletoe species and seven host species. Rates on a leaf area basis. Data are x ± s. Significances refer to analyses of variance. Species Unit transpiration (ml m'" h'*) Sig. Total transpiration (ml h"') Eucalyptus calophylla Amyema miquelii ... 12.48±7.00 34.44+9.74 * 31.87±3.40 Eucalyptus marginata Amyema miquelii ... 38.88 8.10 1.92 .Acacia podalyriaefolia .... Amvema preissii 34.74±26.52 82.80±45.21 * 6.79±3.08 -tm’/a haileyana Amyema preissii 15.90 21.24 — 111.18 Acacia mearnsii Amvema preissii 19.80 140.4 — 35.70 .Acacia decurrens Amvema preissii 29.34±23.40 130.80±42.42 *** 17.72+17.08 Acacia pycnanlha.... Acacia preissii 4.44 58.50 — 50.18 Table 7 Number of seedlings of the mistletoe Amyema preissii which were attached to 1 7 tree species over 60 weeks at Kalamunda Railway Reserve. 1 00 seeds were placed on each tree. Leaves per mistletoe are x ± s. Tree species 8 weeks 28 weeks 60 weeks Leaves/mis- tleloe .Acacia haileyana + 81 1 1 40 Acacia decurrens + 85 10 5 52±38 Acacia mearnsii + .Acacia 97 8 5 72±24 podalyriaefolia + 96 72 72 58±70 Acacia pycnanlha -v .... 95 7 2 6 Dnandra sessilis# 100 27 0 — Eucalyptus astringens. 91 0* 0 — Eucalyptus bolryoides. Eucalyptus 90 8 0 — calophvUaiA Eucalyptus 100 42 0* — camaldulensis 80 26 0* — Eucalyptus cladocalv.x 66 6* 0 — Eucalyptus globulus.... 94 26 0* — Eucalyptus lehmannii. Eucalyptus 95 22 0* — Icucoxylon Eucalyptus 100 36 0* — marginata# 89 59 0 — Eucalyptus robusta 90 51 0 — Nuytsiajloribunda#. . . . 100 78 60 1 ± 1 I Usual host of Amyema preissii at study site. # Species occurs naturally, rest planted by Kalamunda Shire Council. * Flaking of host bark could have contributed to failure of attachment. Discussion Inspection of 20 potential host species showed that Amyema preissii parasitized all and only Acacia species, while Amyema miquelii was almost confined to one of the 1 1 Eucalyptus species available. This apparent host specificity was confirmed by placing seed Amyema preissii on 17 tree species: all gcrminants died except those on Acacia species. The single exception was tenuous growth on Nuyfsia floribunda — an induced example of cpiparasitism. as this species is a root hemiparasite. Seed of Amyema miquelii was unavailable so the ability of this species to parasitize eucalypts in general remains unresolved. In contrast to E. calophylla, it showed almost no affinity for E. marginata (one tree on the entire Reserve) but Amyema miquelii has been recorded on a wide range of eucalypts elsewhere (Barlow 1966). From these considerations, host preferences of the two mistletoe species appear much more important than perching preferences of bird dispersal agents in accounting for the spread of mistletoes (c.f. Lamont and Pcriy 1977, Lament 1983a). If birds were limiting their spread, it could be hypothesised that the larger species and those trees closer to already infected hosts would carry greater loads of mistletoes. In fact, one of the smallest species, Acacia podalyriaefolia. was most heavily infected and percentage establishment in the host-specificity experiment was greatest for this species. Within species, only the largest individuals carried mistletoes. This is clearly a function of plant age rather than bird preferences; establishment of mistletoes on susceptible species is time-dependent. How'ever, trees, boxh within (Table 1) and between (Table 2) species, which were closer to neighbours bearing mistletoes tended themselves to carry a greater mistletoe load. This can be related to the increased probability of receiving seeds from birds after visiting infected trees. In comparison with their hosts, the two mistletoe i species contained larger amounts of water and nutrients (potassium) on a unit weight basis. .Apart from the inert mass of each mistletoe (mean 4.19 kg) to be supported by the host, whole mistletoes contained on average 1.77 1 water and 4-5 g K. These mistletoes also lost considerable amounts of water, whole plants transpiring on average 29.4 ml per daylight hour during mild autumn weather. Despite these drains on the host's resources, only Acacia podalyriaefolia showed obvious signs of retarded grow'th and the mistletoe load could have contributed to the 9% of its individuals which were dead at the study site. Infected trees w'ere in fact larger than uninfected trees because they w'ere older (contrast Lamont and Southall 1982b). Sap flow is maintained by water potential gradients so that it is understandable that the potentials for the two mistletoes were usually much lower than those of their hosts’ (Table 4). This requirement suggests that host specificity may be related to the water potential properties of mistletoe and potential host (Lamont 1982). This possibility is given limited support by the observation that the xylem potential values of non-host Eucalyptus species were usually lower than those of the host species E, calophylla (Table 5). However, this w'ould need to be tested experimentally, as ^ ranged widely for Amyema preissii on different hosts (Table 4) and it still did not parasitize Eucalyptus species with much higher ^ ^ values than recorded for itself. The present data do not give clear support to the role of water relations in determining host specificity. There does not seem to be a simple solution to the control of these mistletoes. Plants that were broken back to the haustorium resprouled readily from submerged tissue in the haustorium. Death is only assured by cutting the host stem beneath the haustorium. an especially difficult task fox Amyema miquelii located up to a height of 20 m on Eucalyptus calophylla. In terms of possible biological control, the White Wax Scale, Gascaraia destructor, was widespread on Amyema preissii. However, plants with the scale did not appear greatly affected (they may even have increased the water and nutrient demand on the host) and it has a wide host range, especially citrus. Occasional plants of A. preissii 24 Journal of the Royal Society of Western Australia, Vol. 68, Part 1 , 1 985. were heavily grazed, presumably by !ar\'ae of the Mistletoe Butterfly, Og}ris amaryllis or a close relative (pers. obs., Atsait 1981). This group has the advantage of monophagy and numbers could be built up for release in the area. However, the mistletoes would recover by resprouiing during any population crashes of the butterfly. Possums {Tiichosurus spp.) are considered a major predator on mistletoes in .Australia (Barlow 1981) and arc still present in Kalamunda. They would provide the most effective control by preventing flowering and fruit set but they arc unacceptable in residential areas. The final avenue is choice of species for future planting programmes. Reduction of suitable hosts would follow the removal of. or planting trees other than Acacia species, especially A. podalyriaefolta. Amyema preissii is not indigenous to the area and was probably introduced via cultivated Acacia species. Alternatively, susceptible species should be planted at least 100 m apart (Table 1). .A particular problem with A. podalyriaefoUa is that it naturalizes readily and soon builds up a vulnerable population of trees. Eucalyptus and other indigenous and exotic genera beside Acacia do not appear to be very prone to mistletoes. Even Eucalyptus calophylla. the dominant tree on the Reserve and the species with longest exposure to mistletoes, only carried an average load of < 2 mistletoes per tree. However, continued thinning of the population by urban development and lack of recruitment of young plants will lead to an increase in the mistletoe load per tree over time (Table 2. Lamoni and Southall 1 982b). Acknowledgemenix . — The Environmcnial Studies Group ai WAIT met expenses as part of its contribution to the Australian Year of the Tree. The Kalamunda Shire C'ouncil is thanked for Us co-operation. Senior plant physiology students participated in much of the work and Chris Gazey and James Grey executed the host specificity experiment. Or. T. M. Roberts and his colleagues are thanked for comments on an earlier draft of the manuscript. References Alsail, P. R. (1981). — Ant-dependent food plant selection by the mistletoe butterfly Ogyris awarv7/i,s’ (Hycaenidae). Oecologia., 48; 60-63. Barlow, B. A. (1966). — A revision of the Loranlhaceae of Australia and New Zealand. Ausl. J. Bat.. 14 ; 421-499. Barlow. B. A. (1981). — The loranthaceous mistletoes in .Australia. In A. Kcast (cd.). Pvologicat Biogeogmphy in Australia. Junk, The Hague., pp. 556-574. Glaizel, G. (1983). — Mineral nutrition and water relations of liemiparasiiic mistletoes; a question of partitioning. Experiments with Lnranthiis europeus on Quercus petraca and Qucrcus rohur Occotogia. 56; 193-201. Hellmulh. E. O. (1971). — Eco-physiological studies on plants in and and semi-arid regions in. Western .Australia. IV. Comparison of the field physiology of the host .\cacia grashyi and its hcmiparasue. Amyema ncsior under optimal and stress conditions. J. F.coL. 59; 351-363. Lamont. B (1982). — Host range and germination requirements of some South Afriean mistletoes. .S’. Afri.J. Set., 78; 41-42. Lamont. B. (1983a). — Germination of mistletoes. In Calder. A. M. and Bernhardt, P. The tiinlogy of Mistletoes, Academic Press, Australia pp, 129-143. Lamoni, B. (1983). — Mineral nutrition of misletoes. In Calder, A. M. and Bernhardt, P. The Biology of Mistletoes. Academic Press. Australia pp. 185-204. Lamoni. B. and Perry. M. (1977). — The effects of light, osmotic potential and atmospheric gases on germination of the mistletoe Amyema preissii. Ann. Hot., 41 : 203-209. Lamoni, B. B. and Southall. K. J. (1982a). — Distribution of mineral nutrients between the mistletoe, .imyerna preissii, and its host, Acacia acuminata. Ann. Bat.. 49: 721-725. Lamoni. B. B. and Southall. K. J. (1982b). — Biology of the mistletoe .Imvemu preissu on road verges and undisturbed vegetation. Search. 13: 87-88. 25 \ Journal oflhc Royal Socicly orWcsiern Australia. Vol. 68. Pan 2. 1986. p. 29-36. The Whitfords Cusp — its geomorphology, stratigraphy and age structure By V. Semeniuk' and D. J. Scarle- ^ 21 Glenmere Road. Warw'ick WA 6024 ^ 108 Dalkeith Road. Ncdlands WA 6009 Manuscript received 18 June 1985: accepted 20 August 1985 Abstract The Whitfords Cusp is a large triangular accretionary promontory situated at the southern end of the Whitfords-Lancelin sector of the inner Rottnesi shelf coast" The subaerial portion of the accrctionary cusp is composed of a diffuse dune terrain, the submarine portion of the cusp is a shallow (scagrass) bank structure, flanked by gently sloping margins that descend to deep water submarine depressions. The main body of the accrctionary cusp abuts a rocky coastline cut into Pleistocene Limestone on the mainand. The cusp is developed leeward of a cluster of rocky prominences that comprise the Marmion Reef and Spearwood ridges. The entire accretionary cusp is underlain by a Holocene sequence of Safety Bay Sand and/or Becher Sand. Radiocarbon analyses together with sealevel indications indicate that the cusp began accumulating c. 7860 C''* yrs BP with sealevel lower than present. Sealevel rose until it reached its present position and stabilised about 5000 C'‘’ yrs BP. Reconstruction of isochrons (age structure) indicates that in the ver\' late Holocene (i.e. c. 1300 yrs BP to the present) the cusp has undergone a major erosional phase. Introduction The coastal environment of southwestern Australia encompassing the inner Roltncst Shelf is composed of Holocene accretionary' sequences and limestone rocky shores. Recently Scarle & Semeniuk (1985) described and classified this coastal environment and established five broad regional sectors each with its own gcomorphology and style of sedimentation. Within this framework. Sector 4. the Whitfords-Lancelin Sector (Fig. 1 ). is composed of a shoreline mostly of limestone rocky shores with isolated intermittent large-scale accrctionary cusps of Holocene sediment. Semeniuk Johnson (1985) described limestone rocky shorc.s that dominate this sector, but to dale there have been no published details of the accretionary components of the Sector 4 system. Semeniuk & Johnson (1982) have described bcach/dune sequences in the Safety Bay Sand previously in this area, and Semeniuk & Scarlc (1985a) have described the broad stratigraphy of the Whitfords area as a framework to a study of groundw'atcr calcrete, but these works do not provide the detail of geomorphology, stratigraphy and age structure provided herein. This paper presents information on the geomorphology, sedimentology, stratigraphy and age structure of the Whitfords Cusp, one of the best developed and largest accretionary cusps in the Whitfords-Lancelin Sector (Fig. 1), so that the area can serve as an example of Holocene accretion in this system. The term accretionary cusp is used in a macroscopic sense to refer to the large scale triangular sandy promontory in the area (see “cuspate foreland'’ in Bates & Jackson, 1980). The term as used here is equivalent to “cuspate sandv foreland” of Bird (1976). The term should not be confused with small scale beach cusps that are developed periodically as a rhythmic feature along a shoreline. The field methods used in this study included (Fig. 2): I) mapping of sediment facies by ground traverses and diver traverses; 2) drilling by reverse circulation air core ( 1 0 sites); 3) angering by a vehicle mounted Gemco rig (4 sites); 4) pit examinations (10 sites); 5) airlift coring in underwater locations (7 sites); 6) levelling of sites relative to AMD and 7) collection of surface samples for laboratoiy analysis (47 sites). The laboratory methods included: I) description of sediment in terms of fabric, texture, composition, colour: 2) aerial photograph interpretation, and 3) sorting of shells from air core material for radiocarbon analysis. The procedure followed in shell sorting is outlined in Searle & Woods (1986). Regional setting The Whitfords Cusp is a Holocene accretionary coastal deposit developed along the modern shoreline of the Swan Coastal Plain. As such it is one of a number of isolated Holocene accrclionar>' cusps developed along the Whitfords-Lancelin Sector of the inner Rotlnest Shelf coast. The nearshore and shoreline zone of this sector is characterised by a variety of features. The nearshore bathymetry to depths of 30 m is characterised by well defined largely submarine shore-parallel limestone rocky ridges, that may also form reefs and islands. These ridges from east to west are termed the Spearwood Ridge, the Marmion Reef Ridge and the 44193-1 Journal of the Royal Society of Western Australia. Vol. 68. I^art 2. 1986. Figure 1. — A. Regional setting of the study area within the sectors of the Rottnest Shelf B. The Whilford Cusp within the Whiiford-Lancelin Sector. 30 Journal of the Royal Socicly of Western Australia, Vol. 68. Part 2. 1 986. Figure 2. — A. Sampling sites and traverses used in this study. B. Facies distribution in the study area. Staggie Reef Ridge (Searle & Semeniuk 1985). The coastline itself in this sector consists largely of a rocky limestone coast and pocket beaches with the discrcte/isolaled dune-topped promontories or accretionary cusps. The subacrial dune terrain of the accretionary' cusp, comprised of fixed and mobile dunes, IS referred to the Quindalup Dunes of Mac.-Vrthur & Bcllany ( 1 960). The oceanographic system along this portion of coast is typical of the regional pattern (Stcedman & Craig 1983. Searle & Semeniuk 1985). The coastal zone is microiidal. In summer the prevailing wave regime is oceanic swell deriving from between west and southwest. This is supplemented by locally generated wind waves developed by scabrcczes. The complex bathymetry' of the nearshore marine enviroment dampens, refracts and diffracts the swell, as well as the locally-generated seas, developing complex convergences and divergences of wave orthogonals. This results in sediment transport and local sites of accumulation in loci of shelter i.c. the main Whitford Cusp. In winter, locally-generated wind waves are a significant influence supplementary to .swell close inshore and during storms. Waves generated by storms approach mainly from northwest and west and are a major influence in sediment transport. The Whitfords Cusp The Whitfords Cusp system incorporates the area shown in Fig. 2B. Its natural boundaries are 1) the contact between Holocene sand and Pleistocene Tamala Limestone which defines the east margin; 2) the junction between Holocene sand and the limestone rocky shores to north and to south; and 3) the Marmion Reef Ridge which forms a sharp contact to west. Essentially the cusp is localised behind the island, reefs and rocky prominences around the Little Island group of the Marmion Reef Ridge. Geomorpholog}' The Whitfords Cusp is composed of a subaerial triangular promontory, termed in this paper an accrelionary cusp, and a submarine extension of this promontory into the marine enviroment as a subaqueous promontory or bank, which has been termed geographically Lai Bank. The subaerial portion of the cusp is composed of a diffuse dune terrain (Fig. 2B) wherein there are fixed steep parabolic dunes, fixed sleep conical dune residuals, vegetated low dunes, mobile parabolic dunes and interdune depressions. MacArthur & Barlle (1980) have mapped various stages of dune development in this area. They subdivided the fixed and mobile dunes of the Whitfords Cusp into a limc-rclatcd geomorphic-soil series of subunits termed Ql, Q2. Q3. 04, with Ql. the oldest, and Q4 the youngest. Essentially all dunes arc in various stages of vegetation and geomorphic degradation with moderate to thin soil cover. The dunes are aligned in a 80“-90“ trend and indicate that much of the terrain is comprised of overlapping parabolic dune blow’outs now largely fixed by vegetation. The submarine promontory, that extends from Mullaloo Point of the Whitfords Cusp and its flanking beaches, to Little Island/the Marmion Reef ridge, is a shallow bank structure remarkably uniform in its depth 31 SAFETY BAY SAND (LITTORAL FACIES) S7 TAMALA LIMESTONE MEDIUM TO FINE QUARTZ/ SKELETAL SAND COARSE TO FINE QUARTZ/ SKELETAL SAND SANDY MUD/MUDDY SAND MUD YELLOW QUARTZ SAND BIOTURBATION SHELL CEMENTED PATCHES/LAYERS SEAGRASS LEAVES. FIBRES AND ROOTS AEOLIANITE LIMESTONE CLASTS AEOLIANITE LIMESTONE Sampled for radiocarbon analysis. Sample numbers are described in Table 2 Figure 3. — A. Stratigraphic profiles across the Whitford Cusp. Location of cross sections are shown on Figure 2A. B. Detailed stratigraphic columns showing the onshore sequence. Journal of ihe Royal Society of Western Australia. Vol. 68. Part 2, 1986. along its cresi-axis. The bank is some 4 km long, 1 km wide and 2-3 m deep, li descends on its north and south margins into submarine depressions some 8-10 m deep. The submarine promontory is largely seagrass vegetated, and in geologic terms would be termed a seagrass bank. The sloping margins of the bank are termed here marginal seagrass bank (Fig. 2B). Scattered through the shallow waters are limestone rocky reefs (e.g. Cow Rocks, Boyinaboat Reef, North Lump) that are part of the Spearwood Ridge. These rocky prominences however are mostly buried by Holocene sediments. Leeward of some reefs there are large scale sand waves 1- 2 km long, up to 1 km wide and with 3-6 m relief and largely vegetation free. The strip of shoreline separating the subaerial and submarine portion of the Whitfords Cusp is composed of beach, beachridges and foredunes wherein the geomorphic/sedimentary units of inshore, swash backshores and aeolian beachridge zones of Semeniuk & Johnson (1982) can be recognised. To north and south, the cusp is bounded by rocky shores cut into Pleistocene Tamala Limestone. Sedimentary facies The Holocene sedimentary facies of the Whitfords Cusp are relatively simple: they are (Table 1 ): 1) dune sand facies, 2) beach sand facies, 3) seagrass bank facies and marginal seagrass bank facies, 4) sand wave facies, 5) depression (or basin) facies, 6) rocky reef facies. Each of the facies may contain several interrelated sediment types (Table 1). The distribution of these facies is shown in Fig. 2B. Table 1 Description of Holocene Facies in Whitfords Cusp Facies Sediments Description Dune facies dune sand soil calcrete cross-layered to structureless to root- structured medium to fine, cream quartz skeletal sand struclurelcs-s to root structured to biolurbaled humic, grey to brown quanz skeletal sand occurs in localised and isolated thin lenses only* Beach facies inshore swash backshore beachridge trough-layered to layered shelly, medium and coarse quartz skeletal sand seaward-inclined. layered (shelly) sand; medium, coarse and fine quart/ skeletal sand layered tn disrupted (shelly) medium, coarse and fine quart/ skeletal sand cross-layered to structureless fine and medium quanz skeletal sand Seagrass bank facies and marginal seagrass bank facies sand and shelly sand muddy sand biolurbaled to structureless to crudely layered (shelly) coarse medium and fine quartz skeletal sand biolurbaled to structureless to crudely layered (shelly) coarse, medium and fine quartz skeletal sand with inteTsiitial mud Sand wave facies sand structureless to laminated fine and medium quartz skeletal sand Depression facies sand and shelly sand medium and fine sand grading to coarse and medium skeletal quartz sand with shell gravel, grading to coarse quartz sand Rocky reef facies sand and shelly sand apron and lenses of medium, coarse and fine quartz, skeletal lithoclast sand locally with shell and lithoclast gravel * See .Semeniuk & Scarle 1985a trending across cusp and truncated by modern shore. These isochrons essentially represent shorelines at the varying age intervals because shells from beach facies were used in the radiocarbon analysis. Figure 5. — Interpreted sealevel history of the Whitford area as determined by data of sealevel indicators and radiocarbon dates. 44193-2 33 Journal ofthc Roval Society of Western Australia. Vol. 68. Ran 2. 1986. Stratigraphy The main Quaternary formations in the area are: • Safety Bav Sand {Passmore 1970. Semeniuk & Searle 1985b)' • Becher Sand (Semeniuk & Searle 1 985b) • Cooloongup Sand (Passmore 1970. Playford ct al. 1976) • Tamala Limestone (Playford el al. 1976) In essence the stratigraphic sequence underlying the Whitfords Cusp is typical of the Bccher Sand/Safely Bay Sand relationships described elsewhere by Searle (1984) and Semeniuk & Searle (1985b). Profiles showing the relationships between the formations are shown in Fig. 3. The Safely Bay Sand is composed of sediment of the dune and beach facies described above. The Becher Sand is composed of sediment of the seagrass bank facies. The Cooloongup Sand is a Pleistocene unit of yellow to orange quartz sand: the Tamala Limestone is a Pleistocene unit of acolianite with marine intercalations. Important features about the stratigraphic relationships are discussed below. The entire Whitfords Cusp and Lai Bank is underlain by a Holocene sequence of sediments that indicate shoaling: seagrass bank sediments are overlain onshore by beach sediments and in turn these are overlain by dune sediments (Fig. 3). The Holocene sequence rests on an unconformity cut into Pleistocene limestone, and along the eastern extremity of the Holocene deposits, the sequence is plastered on and abuts a buried rocky shore (cliff) cut into the limestone. Locally the Holocene sequence rests on Pleistocene Cooloongup Sand. Depressions in the limestone (e.g. the depression between the Spearwood Ridge and the mainland shore) are filled with and levelled by the Becher Sand. The contact between Safely Bay Sand and Becher Sand represents a contact between beach and shoreface sediments and sublitloral seagrass sediments and as such it is a stratigraphic interface (=sealcvel indicator 1) that may be used as a MSL indicator (Searle & Woods 1 986). The modern contact between Safety Bay Sand and Becher Sand is approximately 1.5-1. 7 m below present MSL corresponding to the modern interface between beach/shorefacc and seagrass bank facies. However in older parts of the sequence the contact between Safety Bay and Becher formations is up to 2.7 m below present MSL. Use of the littoral facies (=sealevcl indicator 2) within the Safely Bay Sand (p. 325 Semeniuk & Johnson 1982) provides similar results. The littoral facies of the Safety Bay Sand was deposited at and near MSL. An examination of the stratigraphic profiles in Fig. 3B indicates that for most of the sections the littoral facies accumulated al about present sealevel. In the older portions of the sections however the littoral facies occurs below the present position of MSL. Age of sequence and age structure of cusp Samples of shell and peat were collected from various intervals of the stratigraphic profiles (Fig. 3) and submitted for radiocarbon analysis for dating of the Holocene sequences. The materials used for radiocarbon analyses and the ages they returned are described in Table 2. The radiocarbon results are used here to: 1. confirm the Holocene age of the Whitfords Cusp (Fig. 4A). 2. determine the initiation and history of Holocene sedimentation in this area. 3. determine the rale of accretion of the stratigraphic sequence. 4. determine the age structure of the cusp and its relationship to modem geomorphology (Fig. 4B). and 5. determine the history of sealevel during the Holocene by dating the sealevel indicators. The results from radiocarbon ages have confirmed that the sequence of Bccher Sand and Safety Bay Sand are wholly Holocene (Tables 2 and' 3). Age determinations from the shell and peat in deep portions of site S6, and shell from site S8, indicate that the post- glacial marine transgression had reached this shoreline by 7700-7860 C'"* yrs BI’, and that beach conditions were established al site S9 by 7415 C'-* yrs BP. .Ages relumed from sites S6 and S8 indicate that the bulk of a seagrass bank can shoal from deep water facies to beach fable 2 Description ofmalcrial used for radiocarbon dating Sample no. Core site Lab* no. Depth Formation Type of material • A/nouni (s) Why sampled ** - l.tfc C'Urs with C * correciion 1 2 SI GX 10637 6- 7 m SafetN Bas Sand Donax 10 1 1 345 - 1 70 S2 CiX 10675 9.9.5 m Safets Bay Sand Largely Dimax (some Brachtdontes) 24 1 .3485 - 150 3 S3 S5 CiX 10638 5-6m Safetv Bav Sand Oonax 14 1 AQ1 5 * *>70 4 GX 10676 6*7. 5m Safety Bav Sand Donax 14 1 3790 . 145 5 S6 (iX 10677 3 -4 m Safety Bay .Sand Donax • (tlycymeris. Brachukmu's 10 1 3870 * 205 6 S6 GX 10678 14- 1 5m Becher Sand BrachidonU’s 7 -> 7770 '175 7 8 S6 CiX 10679 14- 15m Becher Sand Seagrass peat 5 3 7295 • 130 S7 GXl 1 101 5-6m Gravelly shelly sand at unconformity Mixed molluscs, mainly rocky shore assemblage 13 4 5585 ± 170 9 S8 GX 10680 6. 5-7, 5m Safetv Bav Sand honux 20 1 5115 * 1 A5 10 S8 CfX 1 1 100 I7.|8m Becher Sand Mixed gastropods Thuloiia. BhasHinvUa etc. 14 -1 7860 * 230 1 1 S^^ GXl 1099 9-1 Im Safety Bay Sand Donax and mixed molluses 5.5 1 7415 • 360 * Laboratory No. lor radiocarbon analysis Gcochron **1 Division. Krueger Enterprises Inc. ■> < XRD and thin section analyses from all these materials show no diageneiic alteration of shell 4. to determine age of beach facies to determine age of lower part of Becher Sand which in conjunction with age of beach facies provides data on rate of shoaling supplemenlar\ material to complement and confirm age of shell from same hon/on age of shell deposit to date overK ing beach facies 34 Journal oflhc Royal Society of Western Australia. Vol. 68. Part 2. 1986. Table 3 Data on scalevel indicators Sealevel indicator 1 Sealevel indicator 2 Site* No. Sample No. used for radiometric dating Age of indicator C'* yrs BP C ’ corrected Beach sand contact with Becher Sand Position of sealevel at lime of deposition Littoral facies Safety Bay sand Position of sealevel at time of deposition SI 1 1345 Shelly coarse and medium sand overlying a sediment with a fine fraction present level Shelly coarse and medium sand present level S2 2 3485 Shelly coarse and medium sand overlying a sediment with a fine fraction present level Shelly coarse and medium sand present level S3 3 6915 Shelly coarse and medium sand overlying a sediment with a fine fraction n.8m below present Shelly coarse and medium sand c. 1 .Om below present S5 4 3790 Shelly coarse and medium sand overlying a sediment with a fine fraction present level Shelly coarse and medium sand present level S6 5 3870 Shelly coarse and medium sand overlying a sediment wiih a fine fraction present level Shelly coarse and medium sand present level S7 8 5585 Not present Shelly coarse and medium sand overlying gravel and shell on unconformity c. present level S8 9 5115 .Shelly coarse and medium sand overlying a sediment with a fine fraction present level Shelly coarse and medium sand present level S9 1 1 7415 Shelly coarse and medium sand o\cr]ymg a sediment with a fine fraction 1 -Om below present level Shelly coarse and medium sand c. LOm below present * see Table 2 and Fig. 3 for location and depths. facies in c. 2700-3900 C'-* yrs (i.e. 7860-51 15 and 7700- 3870 respectively). A map showing ages of the shells from the beach facies at various sites indicates that the age structure of the Whitfords Cusp is internally consistent (Fig. 4B). However the interpreted isochron distributions are truncated by the modem shoreline, indicating that the cusp is in a major erosional phase. The dating of sealevel indicators at the various sites enables a reconstruction of sealevel history for the interval of the Holocene c. 7860 C’’ yrs BP to the present. (See Searle & Woods, 1986 and Semeniuk & Scarle, 1986 for discussion of .sealevel indicators in accrctionary sequences.) The critical data pertaining to the two types of sealevel indicators, their ages and their stratigraphic level or position relative to present MSL are presented in Table 3. The stratigraphic profile in Fig. 3 clearly illustrates that earlier in the Holocene some sedimentary units were deposited with MSL slightly below present — for instance the Becher SancFSafety Bay Sand contact at sites S3 and S9 is some 2-2.5 m below present MSL. On the other hand the same interface is 1.5- 1. 7 m below present MSL at sites SI, S2, S5, S6. and S8. The modern Becher Sand/Safety Bay Sand contact occurs as an interface 1.5-1. 7m below present MSL. The littoral facies of the Safely Bay Sand shows a similar pattern. The facies occurs at about MSL for sites SI. S2, S5, S6, S7 and S8, but occurs below MSL at sites S3 and S9. The evidence above indicates that there was deposition of sedimentary' units earlier in the Holocene when sealevel stood c. 1 m below present. Age deieiminaiions of these units shows all sequences younger than c. 5000 €'■* yrs BP as having formed with sealevel at about present position. Sequences older than 6900 C'"* yrs BP were deposited with sealevel c. 1 m below present. The interpreted sealevel curve derived from the radiocarbon age data and levels of MSL indicators is shown in Fig. 5. Developmental history of cusp The post-glacial marine transgression reached the shore of the Whitfords area some 8000 yrs BP. The initial stages of this marine incursion resulted in development of rocky shores as marine erosion incised into the Tamala Limestone terrain. Shore deposition began in this area c. 7415 C'-* yrs BP with sealevel I m below present and steadily rising. About 5000 C'** yrs BP sealevel had reached approximately its present position but by this stage a significant volume of the cusp had accumulated in the form of a seagrass bank capped by beach and dune sediments. The accumulation was developed behind the sheltered loci of the cluster of barrier islands, reefs and ridges of the Marmion Reef Ridge centred on Lillie Island. By c. 1300 C'^ yrs BP the Whitford Cusp had accreted to its maximum width as preserved today. However in the very late Holocene the cusp has gone through a major erosional phase where the shoreline has retreated and incised into the Holocene deposits, markedly truncating the time planes (isochrons). The north side of the cusp has been cut back nearly to the 5000 C'*’ yr isochron; the south side has been cut back nearly to the 3780 C’** yr isochron; the tip of the cusp has been cut back to the 1345 yr isochron. The resposilory of the eroded material is not known at present but it may either have moved out of the area as a shoreline ribbon, or it may have been lost into the adjoining depressions. Discussion and conclusions The Whitfords Cusp accretionary system is broadly similar to those described by Scarle (1984) and Searic & Semeniuk (1985) in the Cape Bouvard-Trigg Is. sector of the Rottnesl shelf in that the Holocene sequence consists 35 Journal oflhe Royal Society of Western Australia. Vol. 68. Part 2. 1 986. of BccherSand underlying Safety Bay Sand (Semeniuk & Searlc 1985b). However the Whiifords Cusp differs from the Cape Bouvard-Trigg Sector cusps in a number of aspects. Firstly the Whitfords Cusp is a relatively isolated accretionary' cusp developed in the lee of an island/rocky reef area whereas the accretion in the Cape Bouvard- Trigg Sector locally has resulted in a series of adjoining cusps that have coalesced to form a broad prograded plain (e.g. Rockingham plain). Secondly^ whereas the prograded plain at Rockingham still has clearly preserved surface beachridge trends (Fairbridge 1950. Seddon 1972, Woods &. Searle 1983), the accretionary beachridge growth lines of the Whiifords Cusp have long been erased and overprinted by landward migrating parabolic dune blowouts. In the Rockingham Plain area the surface gcomorphology (beachridges) trends were used by Woods & Searlc 1983 and Searle & Woods 1986 to determine the history' and age structure of the prograded plain, but in the Whiifords Cusp the age structure and growth trends (Fig. 4) can only be determined from subsurface information. The extreme truncation of growth trends of the Whitfords Cusp by the modern shoreline (Fig. 4B) also contrasts with the history of accretionary cusps in the Cape Bouvard-Trigg coastal sector. These latter accretionary cusps generally are in geomorphic equilibrium with the growth lines/age structure and indicate that accretion is still proceeding. The Whitfords Cusp, however, has ceased prograding and is now in a major erosional phase. The final aspect to emerge from the Whitfords Cusp area is the sealevel history as determined by sealevel indicators and the radiocarbon ages of selected stratigraphic intervals. Searle & Woods (1986) recently discussed the significance of the differences between sealevel curves determined from accreiionary sequences as compared with those determined from rocky shores. The sealevel history curve detennined from the Whitfords Cusp however not only is markedly different to curves derived from rocky shores in the region (Fairbridge 1961, Playford 1977) but it is also different from those derived from other accretionary shores such as Leschenauli Peninsula (Semeniuk 1985) and Rockingham Plain (Searle & Woods 1986). This marked variation in sealevel history along a relatively short segment (170 km) of the Western Australia coastline is attributable to tectonic influences, a factor also raised by Playford ( 1 977) to account for discrepancies between the sealevel history at Rolinest Island and Fairbridge (1961). The significance of the variable sealevel history along this coastline is discussed further in Semeniuk 8 l Searle 1986. Acknowledgements — Funding for core sites SI*S10 and for radiocarbon analysis of sample numbers 1-6 and 8 was provided by the Public Works Department of W.A. as part of the environmental study by LeProvost Semeniuk & Chalmer into the Hil!ar>'s/Sorrenio Marina. References Bates. J. A. and Jackson, R. L. (1980). — Glossary of Geology. 2nd Ed. AM. Geol. Instil. 75 )p. Bird, E. C. F. (1976). — Coasts. 2nd Ed. ANU Press. Canberra. Fairbridge. R. W. (1950). — Geology and geomorphology of Point Peron. Western Australia. Jnl. Koy. Soc. West. .4usir.. 34: 35-72. Fairbridge. R. W, (1961). — Eustalic changes in sealevel. /n PHysics and Chemistry of the Earth. L. H. .Ahrens, F. Press, K. Rankema and S. K. Runcorn (cds). 4. 99-185, Pergamon. Oxford. MacArthur, W. M, and Bartle, G. A. (1980). — Landform.s and .soils as an aid to Urban planning in the Penh Metropolitan Northwest Corridor, Western -Australia. Land Resources Management Series No. 5.. C.S.I.R.O., Australia. MacArthur. W. M.. and Beiienay. E. (1960). — The development and distribution of the soils of the Swan Coastal Plain. W.A. Soil Pub. No. 16., C.S.I.R.O., Melbourne, Passmore. J. R. (1970). — Shallow coastal aquifers in the Rockingham district. Western .Australia. Water. Res. Found. Australia Bull., 18. 83p. Playford. P. E. (1977). — Part 1, Geology and groundwater potential in Playford. P. E. and Leech. R. E. J.. Geology and hydrology of Rotinest Island. Geol. Surw West. .‘(wj/.,Repi 6; 1-53. Playford. P. E.. Cockbain. A. E. and Low. G. H, (1976). — Geology of the Perth Basin. Western Australia. UW/. Aust. Geol. Surv. Bull.. 124: 31 Ip. Searle, D. J, (1984). — A sedimentation model of the Cape Bouvard to Trigg Island sector of the Roitnesi Shelf. Western Australia. Unpubl. Ph.D. Thesis. Univ. West. .Aust. Searlc. D. J. and Semeniuk. V. (19851. — The natural sectors oflhe inner Rollne.si Shelf coast adjoining the Swan Coastal Plain. Jnl. Roy. .Vw, U’. 67; Searle. D. J., and Woods, P. (1986). — Detailed documentation of a raised Holocene sealevel record, west coast. Western Australia. (Quat. Res., in press). Seddon. G. ( 1 972). — Sense of Place. University of Western Australia Press, Ncdlands. 214p. Semeniuk, V'. (1985). — The Age Structure of a Holocene Barrier Dune System and its implication for sealevel history reconstructions in Southwestern .Australia. MarineGeoi, 67, 197-212. Semeniuk. V.. and Johnson. D. P (1982). — Recent and Pleistocene bcach/dune sequences. Western Australia. Sediment. Geol.. 32; 301-328, Semeniuk. V. and Johnson. D. P. (1985). — Modern and Pleistocene rocky shore sequences along carbonate coastlines, southwestern .Australia. Sediment. Geol . 44- 22 5-26 1 . Semeniuk. V,. and Searle, D. .1. (1985a). — Distribution of caicreio in Holocene coastal sand in relationship to climate, southwestern Australia../ Sed. Petrol., 55; 86-95. Semeniuk. V. and Searle. D, J. (1985b). — The Bccher Sand, a new stratigraphic unit for Holocene sequences of the Penh Basin. JnL Roy. Soc. IF. 67; ScmcMiiuk. V. and Searle. 1). .1. (1986). — Variability of Holocene sealevel historv along the southwestern coast of Australia — evidence for the effect of significant local iccionisni. Marine Geol. in press. Steedman, R. K.. and C’raig. P. D. (1983). — Wind-driven circulation of ('ockbum Sound.-Jw-yrr J. Afar Res..M: 187-212. Woods. P., and Searle. D. J, (1983). — Radiocarbon Dating and Holocene History- of the Becher/Rockingham Beach Ridge Plain. West Coast. Western .Australia. Search. 14; 44-46. 36 Journal of the Royal Society of Western Australia. Vol. 68. Part 2. 1986. p. 37-50. A Biology of the desert fringe Presidential address — 1984 by S. J. J. F. Davies Royal Australasian Ornithologists Union. 2 1 Oladstone Street. Moonee Ponds. Victoria. 3039. Abstract Work at Mileura Station. Western Australia, from 1 959 to 1 983 has led to the description of some of the strategies used by plants and animals to survive in this arid area. The low rainfall is concentrated into the creek systems that cover only 10 percent of the land surface. Regular plant production takes place in the creeks so that annuals and perennials both show seasonality to which the animals' life cycles appear to be geared. In years of heavy rainfall the whole land surface is productive and animals breed abundantly. In years of low' rainfall only enough plant food is produced to enable animals to survive and few or none breed. Animal populations survive in the region by a capacity for nomadism or an ability for a few individuals to survive in favoured sites during dry' times coupled with an ability to reproduce rapidly when conditions are good. In addition to physiological and behavioural adaptations to the arid conditions, examples arc presented to suggest that the longevity of animals is determined by the frequency with which years occur when breeding and recruitment arc successful. Longevity must be longer than the longest interval between years in which the resources each species needs are produced abundantly. Populations in which no individuals can live long enough to survive from one productive season to the next will soon be eliminated. Introduction Australia is an arid continent. Of its land surface 66% receives a mean annual precipitation of less than 500 mm (Nix 1982). Most of this land is not bare rock or sand as are some arid lands in Africa. Asia and America, but vegetated with perennial plants. The trees and shrubs of the arid zone merge with those of mesic parts of the continent in a wide zone as a mosaic of arid and mesic plant communities. This zone is the desert fringe. In this paper 1 review the results of work since 1959 at Mileura Station (20“ 22’S; 1 1 T 20'E) in the pastoral area of Western Australia, comparing the results with those from deserts to the north and east and the woodlands to the south and west. Much of the desert fringe in Western Australia is occupied by sheep and cattle stations where the animals arc kept in large paddocks and graze on uncleared native vegetation (rangeland). The effects of the associated range management are considerable, although the appearance of the landscape is little altered. To assess the influence of the pastoral industry on the organisms that live in the desert fringe detailed, almost microscale, studies are needed. Broad scale surveys and superficial assc.ssmenls provide only a hazy outline of the ways organisms have responded to changes instituted by Europeans, some dating back to the 1 860's. For this reason I shall single out those organisms about which much is known, examine how changes have affected their populations and draw some generalizations together from these examples. The environment of a sample of the desert fringe, incuding Mileura. was described in the report of a survey by Mabbutt el al. (1963). Reviews of some of the work done on Mileura have been published by Davies (1968, 1973, 1975, 1976c). Mott (1979) and Watson and Perry (1981). In addition useful reviews of aspects of the biology of arid Australia have been published by Serventy (1971). Frith ( 1 976). Keast (1981). Barker and Greenslade ( 1 982) and Harrington el al. ( 1 984). The environment (ieology. soil and land elassijicalion Mileura Station is 1 50 km north of Cue and about 800 km north-east of Perth. It is a sheep station of approximately 282 000 ha. The geology of the area (Mabbutt 1963) is dominated by rocks of the Archaean Shield, one of the most ancient land surfaces of the earth. The rocks are much weathered granite and gneiss with ranges of metamorphic origin, the Jack Hills, lying across the northern end of the lease. The landscape is flat, dominated by the Pindabarn Creek that runs north- north-west through the centre of the property turning west-south-west at the foot of the Jack Hills (Figure 1). The Pindabarn is a tributary of the Murchison River. Mabbutt et al. ( 1 963) mapped the land systems (areas of a recurring pattern of soil, topography and vegetation) of Mileura, separating the main creek and its flood plains with soils of dcpositional origin from the rocky hills where active erosion is taking place. Figure I (from Davies and Walsh 1979) shows the distribution of the Sherwood and Belele land systems. The Sherwood land system is the main one in the highlands. Although erosion is taking place there the lime scale is a long one. An aboriginal rock shelter beneath an apparently actively eroding escarpment (breakaway) in the Sherwood Land System Belele Land System Fence line • Mill A Quadrat site ■ Homestead Rock hill Watercourse Figure 1. Map of Mileura Station, Cue, showing the distribution of the fences, watering points, paddocks, main pools, watercourses and land systems including the mam hills (After Davies and Walsh 1979). Journal ofthc Royal Socicly of Western Australia. Vol. 68. Part 2. 1986. Sherwood land system on Mileura contained charcoal 50 cm below the present floor that was carbon dated to 1340 ( ^/-lOO years) years BP (Davies et al. 1977). The rale of change of the landscape is therefore slow. The Beiele land system lies between the highlands and the deposilional soils of the creek. Its red. sandy loams slope gently towards the creek and are vegetated with perennial shrubs as well as annual grasses and herbs. The creek system has parts where erosion has been severe so that much of the top soil has gone, the Ero land system, and other parts where deep clays have been deposited, the Mileura and Berringarra land systems. In these deposilional soils calcium salts have accumulated by evaporation, sometimes compacting into limestones but always ensuring that the soils have a high ion content. Small stands of perennial chenopods. now greatly degraded, grew on these deposilional soils. H eather systems The climate of Mileura is dominated by two weather systems separated by the anlicyclonic belt. In the summer (Novembcr-March) this belt moves south and Mileura is under the influence of a monsoonal^ depression with its origin in the tropics. At that lime ot year tropical cyclones or rain-bearing depressions sometimes reach the station, supplementing the rainfall from locally induced thunderstorms with downpours of 100 or more mm. In winter (May-August) the anticyclonic belt moves north and the station receives rain from depressions of southern origin that follow the anticyclonic belt north. The distance of movement of this belt varies from year to year and although Mileura can receive heavy falls of rain in summer and winter, these do not occur every year. Spring (September and October) and autumn (April) are short, usually dry. seasons of transition. Rainfall Rainfall figures for Mileura homestead, in the centre of the property, are available from 1907. The mean annual rainfall for Mileura is 1 98 mm (.Arnold 1963). Such a figure gives no indication of annual or seasonal variability. The importance of interactions between rainfall events and each of ambient temperature and topography can best be appreciated if the rainfall data 100 if) 5 50 < LU > 0 WINTER n SUMMER ■ 25 50 WET PERIOD EXCEEDING (MM) 75 Figure 2. — The rainfall of Mileura showing the per cent of years in which falls in winter (May-August) and summer (Novembcr-March) exceeded the amounts shown along the horizontal axis (After Davies 1 968). are presented as a percentage of years m which rainfall events exceeded precipitation totals of different amounts, analysed separately for summer and winter. In Figure 2 the vertical axis .show's the percentage of years in which at least one wet period exceeded the total precipitation shown along the horizontal axis. Arnold (1963) has defined a wet period as a period of rainy weather that is terminated by more than two dry' days because the soil surface usually remains saturated for as long as that after rain. The figure shows that although large falls of rain arc rare in summer and winter, falls of 25 mm or more were recorded in 65% of the summers and 63% of the winters. Falls of 15 mm or more were recorded in 83% of the summers and 90% of the winters. If such light falls could be shown to be useful to organisms living in the desert fringe the pattern of rainfall takes on a regularity that is usually denied to it (Davies 1968). So little rain is recorded in spring and autumn that it may generally be ignored. Temperature The annual temperature regime at Mileura resembles that at Meekatharra 140 km to the east, where records have been kept for longer. The area has four seasons, each of a different length. Summer when temperatures are high and fairly stable; autumn, a short season when temperatures fall rapidly; winter when temperatures are low and fairly stable; spring, a season of rapidly rising temperatures. Much w'ork has emphasised the influence of the high temperatures of arid lands on the organisms dwelling there (eg. Davies 1982, Dawson 1976). At Mileura I was able to examine the potential both high and low temperatures had to cause stress on organisms living there. Figure 3 illustrates the frequency distribution of temperature thresholds above and below- selected levels (above 33“ and 37“ and below 10“ and 1 5X'). They were recorded over four years at a site on Mileura Station. Records w'ere made on a recording theremohydrograph in a standard Stevenson Screen adjacent to the headwaters of a creek near a granite monolith. It is difticull to determine above and below what threshold temperatures heat and cold stress respectively begin to operate on animals. Other observations at Mileura. quoted by GrifTiihs (1968). showed that rock shelters and caves used by animals ranged in temperature from 15“ to 33'’C through the year; whereas the outside temperature ranged from -T to 48“C. The range 15“ to 3.3“C is assumed to include the comfort range of many of the animals of the area although it may not coincide precisely with the ihermoneutral zone of some of them. It is apparent that using the above criteria there are longer and more frequent periods of cold .stress than there arc of heat stress. It is not biologically meaningful to make a direct statistical comparison between these frequency distributions, because even one long period of stress can be fatal for the animal or so w'eaken it that it is unable to recover in subsequent less stressful conditions. In this regard it should be noted that although no period exceeding 33“C lasted more than 24 hours. 38 periods when the temperature remained below 1 5“C lasted more than 24 hours. An integration of the curves above and below the threshold temperatures might have been a better measure but was not practical in this case, and the figure makes the point sufficiently clear that there were longer periods of cold stress than heat stress at this site on Mileura in the years 1969-73. 39 Below lO'c Below 15 'c Above 33*c Above 37*c Journal of the Royal Society of Western Australia. Vol. 68. Part 2. 1986. 4C 1970 1971 1972 1973 a rfL c£h P 1 r _□ Hours duration Figure 3. — The frequency of periods of difl'crcnl duration when the temperature was (a) above (b) above 33°C. (c) below 15“C, (d) below 10°C — in each of the years 1970. 1971, 1972 and 1973 at a site on Mileura Station. Western Australia (After Davies 1975). Other measurements have shown that the yearly maxima and minima are greater in the creeks (Mileura/Bcrringarra Land Systems) than in the hills (Sherwood Land System). As a result the climate of the hills is milder than that of the creeks. Rainfall and temperature Another aspect of temperature concerns its interaction with rainfall and evaporation parameters. In summer, temperatures are usually high just before rain falls. Once the ground is soaked evaporative cooling brings the temperature down ver>' rapidly so that the ambient temperatures in the days after rain are substantially lower than the mean for the summer. The same is true of winter. Mott (1972a) measured these effects at Mileura and Figure 4, taken from his work, illustrates them. The effect of a fall of rain is thus not just to raise the moisture content of the soil but to reduce ambient temperature as well, a consequence often overlooked in descriptions of the desert environment (Davies 1976c). Topography The flat appearance of the Mileura landscape is deceptive. It is dissected by small creeks and watercourses (Figure 1) each of which drain water from the highlands (Sherwood Land System; see Geology, Soils and Land Classification above) to the main creek. In the highlands relicts of the old land surface remain as breakaways, rising 1 5-30 m above the surface of the plain. In these breakaways are small cliffs, ravines and caves that provide shelter for many animals and some plants, that can thus avoid many of the extremes of ambient temperature. The watercourses that flow away from these highlands begin as small but deeply incised channels that soon branch, forming small floodplains Figure 4. — Soil surface temperatures at Mileura after rainfall in winter (a) and summer (b). Data obtained with a Theiss mercury in steel thermograph at 0.5 cm below the ground surface (After Mott 1972a). 40 Journal of the Royal Society of Western Australia. Vol. 68. I^arl 2. 1986. between the branches. Later the channels become indistinct and the typical ‘wash’ of the mulga zone forms, with few channels and a wide flood plain. The watercourse eventually spills into a larger creek and the cycle of topographical change is repeated until the main ptoliie Land System Shprwood Eio Bpi ’ m^rra beieie Vegetation A vtct A vict A vict A leir A tetr A teir A D'un A pfun A prun t Iras S spic t fras S spic E tras Cassia S spin Cassia S spm Cassia H lore H lore Ramtali needed to Ibw Ma« Temperature ,, ^ Min Jan 37 0 July 8 0 mm 41 i 20 Figure 5. — Diagram illustrating the cyclical changes in the structure and vegetation of creek systems at Mileura Station. Western Australia. Abbreviations; A — Acacia, Cassia — C. desolata and C helmsir. E. fra — Eremophila fraseri: H. lore — Hakea suberea. prun — pruinocarpa, Sant — Sanialum spicatum\ S. spin — Scaevola spinescens: tetr — leiragonophylla, vict — v/c/onac (After Davies 1973). Pindabarn Creek is reached. Davies (1973) has described the vegetationa! changes that arc associated with this lopogragaphical sequence in more detail. They are illustrated in Figure 5. As each stage in the creek’s development changes in the course of each cycle, so the plants change. Because the creeks in the first cycle are small the area of each associated vegetation is small. The extent of pasture available for animals is therefore larger with each successive cycle down the watercourse system. The creeks, watercourses and floodplains into which the water runs after rainfall occupy about 10% of the land surface (Mabbull cf al. 1963). Water penetrates the sands and sandy loams of Mileura rapidly, but is much slower penetrating the clays of the main creek. Penetration at all sites is curtailed by a. hard pan the depth of which varies from 20 cm to over 1 m. High moisture contents arc often just above the hard pan. The surface layers of the soil need to be thoroughly saturated before water starts to run off into the watercourses. Observations at Alice Springs on soils similar to those of Mileura show that run-off started there after 15 mm of rain had fallen (Slaiyer 1961). Rainfall and topography It follows that falls of rain of greater than 15 mm within one wet period lead to run-off of surplus water into the creeks at Mileura. The redistribution of water after rainfall is of great significance to the biology of the desert fringe. As a consequence of it 10% of the land surface, the creeks, watercourses and floodplains, receive p E SI "a Q 0) 1000 _ r Rainfall I - 42 mm II = 5 mm ^ ' Rainfall A - 28mm B = 3 mm C = 5 mm Rainfall 1 = 15 mm 2=6 mm 3=6 mm 4=5 mm 5=11 mm ' T 50 a 17-7-70 “1 b T C “1 ‘ 60 Days Figure 6. — Water depth curves recorded at a site in a creek bed on Mileura Station, Western Australia: (a) changes in water depth in May, June and July, 1970, after 42 mm of rain; (b) changes in water depth in February. March and April, 1971, after 28 mm of rain; (c) changes in water depth in July and August 1973, after 15 mm of rain (After Davies 1975). 41 Journal ofthe Royal Society of \Veslcrn Australia. Vol. 68. Part 2. 1986. a greatly enhanced effective rainfall. Figure 6 shows some observations made tn a creek bed at Mileura. Following rainfalls the stream began to flow over what had been a dry sandy creek bed. Small falls of rain led to flows of water in the creek 12 to 40 times as deep as was recorded by standard rainfall measuring equipment. Much of this redistributed water flows away down the creeks but about one third remains, stored in the soils of the creek bed and adjacent flood plains for months after the rainfall event. There it can be lapped by the roots of plants, enabling the plants to grow in these watercourses long after the moisture reserves on the other 90% of the landscape have disappeared. Because wet periods exceeding 1 5 mm arc recorded from a high percentage of the summers and winters, any organism that depends on the watercourses (run-on sites or mesic areas) can reproduce in most years, the productivity of the episode depending on the amount of water and the distance out from the creek to which the flood spreads. Soil moisliirv and rain fall In 1980 it was possible to examine changes in soil moisture following rainfall events over a full year. Eleven sampling stations were laid out across the Pindabarn Creek, extending from the highlands to the east, across the creek, over the highlands to the west and some way down into the watershed of the next creek. The moisture contents of soils at six different depths, expressed as % dry weight for each site, are shown in Figure 7. displayed to show' the measurement’s relation to rainfall episodes. The efl'ective penetration of rain that fell in winter was greater than that of rain that fell in summer and the rate of penetration in the clays of the main creek slower than that in the sandy loams of the Belele and Sherwood land systems. Once the soil had been wet moisture was retained longer in the clays than in the sandy soils, but the increase in soil moisture was much quicker in the soils of the highlands. A consequence of these effects was that the small creeks in the highlands responded rapidly to rain, plants germinating and establishing after light falls but rapidly dying off as the moisture drained and evaporated away. In the main creek the soils were not wet sufficiently to ensure germination and establishment by light falls, but when heavy rain fell and soaked the soil, the annuals grew proliflcally and the shrubs flowered and fruited well. Plants Gennination and establishment Most of the data about germination relates to annual plants. Mott {1972a and b, 1973) showed that of those examined the grass and herbs of Mileura differed in their temperature and soil moisture requirements. For example, the grass Aristida contona germinated onlv after it had been exposed to the equivalent of summer temperatures for some weeks. For germination to take place the seed bed had to be saturated (at field capacity) for at least 24 hr and at a temperature above 22-23°C. The daisy Helipterum crasspediodes needed a period of dormancy, but required 2 to 3 days of saturated seed bed at 13-15°C. As a result the grass germinated only after rain in summer and the daisy only after rain in winter. The germination of mulga Acacia aneura has been studied by Preece (1971). He showed that the first germination could take place if a wet period exceeding 25 mm occured in the summer after the seed was set. Little seed germinated then, much being “hard” seed that germinated after rainfall in subsequent summers. Working in inland New South Wales Preece calculated that mulga could regenerate there once every nine years. Favourable sequences of rainfall episodes may be rarer than this in Western Australia because few examples of w idespread shrub regeneration can be seen on Mileura. Nevertheless regeneration of shrubs is taking place. Davies and Walsh (1979) showed that, despite commercial stocking with merino sheep, more individuals of 24 out of 30 species of shrub were located in sixty fixed quadrats in 1976 compared with 1967. One species Kremophila fraseri regenerated better from seed at high stocking rates and two others Solanum ashbyae and Acacia (etragonophylla better at low slocking rates. Acacia aneura had regenerated slowly but its rate of regeneration did not correlate with stocking rale. Such trends could lead to changes in species diversity in the area. The annual floras A preliminary list of plants of Mileura is contained in Davies (1970). The annual flora of the region is ' dominated in winter by daisies (Asteraceae) and in . summer by grasses (Poaceae). The same species do not germinate and establish each year, presumably because each species has specific requirements that are only met occasionaly. Some species can always be found, for example //. crasspediodes and A. contorta. although much more abundantly in some years than others, whereas other species seem to be absent altogether in | some years, eg. some Calandrinia species. Mott (1979) showed that two common annuals H. crasspediodes and A. contorta had seeds most of which j germinated the year after they were produced. In i addition the viability of any stored seed was greatly reduced in the second year compared with the first. The significance of these observations is that these two , common species appear to have evolved a life history' strategy that depends upon annual reproduction. The success of these plants attests the success ofthe strategy. It presumably means that the rainfall is sufficiently regular, both in summer and winter, for redistribulon of water to ensure that the conditions the annuals need for germination and establishment occur each year. Both species arc widespread on Mileura. .As neither have particularly good mechanisms to ensure wind or other dispersal of seeds this wide distribution must depend on locally produced seed. In exceptionally wet years the grass and herbs grow' luxuriantly on the flat lands between the watercourses, but grow each year m the watercourses (run-on sites). The perennial flora Acacia is the dominant shrub genus at Mileura. with Eremophila and Cassia shrubs forming a sub-storey beneath the tree-like shrubs. The densest and tallest stands of shrubs are along the watercourses. Nowhere on Mileura do they approach the size and luxuriance of the mulga A. aneura woodlands of southern Queensland. 42 ll:i: 80 Site 0 n o 50 - 14 ;i: 80 Sampling 13 ;iv;'80 date 11 :vii :’80 Moisture content ( % Wt) O = < 0-01 % 24 :x:'80 -1 lO Figure 7. — Records of soil moisture content (% dry weight) from six different depths at eleven different sites in January (2 readings). April. July and October 1980 and January 1981 . Each histogram represents the mean of two sets of measurements, using a gravimetric method of assessing soil moisture content. The eleven sites traversed a 16 km route transecting the Pindabarn C reek on Milcura Station, Western Australia. The vertical a.xis shows, for each site, a vertical scale giving the depths in cm at which the samples were taken. The horizontal axis shows, for each sampling date, the % dry- weight of water in the samples. Falls of ram at Milcura 1979-81: June 1979; 5.5 mm; Aug. 1979: 1 mm; Dec. 1979: 8 mm; 12.1.80; 36 mm; 15.1.80; 1.6 mm; 28.1.80: 3 mm; 10.2.80: 26 mm; 13.2.80: 2.2 mm; 9.4.80: 4 mm: 20.4.80; 18 mm; May 1980: 33.2 mm; June 1980: 107.3 mm; 10.7.80; 4 mm; 18.7.80: 53 mm; 26.7.80; 4 mm: 26.8.80: 4 mm: 1 8. 10.80: 7 mm; 1 1. 1 .8 1 : 4.5 mm. The main creek is covered by Stations 6. 7. 8 and 9. Journal oflhc Royal Society of Western Australia. Vol. 68. Rail 2. 1986 The flowering and fruiting season of many of the shrubs are remarkably regular (Davies 1976a). Figure 8 illustrates the flowering and fruiting phenology ot 2^ common shrubs on Mileura. Notice the brief flowering periods of A. pndnocarpa and .-F vicloriac. A. aneura flowered over many months but only the flowers of January and February set seed (Davies 1968). Eremophila fraseri was one of the few shrubs that flowered almost throughout the year. For most there was a distinct flowering period and a time of year, usually spring and summer, when fruit was produced. The regularity was made possible by the redistribution and concentration of water. Most fruit was produced in the watercourses. 80 15 HV -V 1 n V J 11 ,0 I I I I I I -I L FJDM FJOM I960 1961 I960 1961 21 22 FJDM I960 1961 Figure 8. — The fluNvering and fruuing phenology of 22 species of shrubs on Mileura Siaiion, Western Australia. Each hisiograni shows the proportion of the sample tn tlowcr each month (above the mid-line), and the proportion of the sample fruiting each month (below the mid- line). The months marked are February. July. December and May. 1; Muircana conw.xa: 2 Sawvola spmcscens: 3. .h'ucta udsiirgens: 4: I aneura: 5: .1, craspedocarpa: 6: -I. eufhherfsonii: 7; 1. kempeana: 8; pruinocarpa: 9: 1, selero.sperma: 10: ,J. lerraaonophylla: 1 1; A. victonae: 12; Acacia sp. nov. HA 251274; 1.3: Acacia sp. nov. HA 251275; 14: Cassia desniata: 1 5; C hdmsu: 1 6: Eremophila frascri: 1 7: E. frcclindk 18: E. Icucophylla: 19: Hakca torca: 20; Sanlalum lanccotalum: 21: S. spicalum: 22: Solanum ashbyac. Abbreviation: HA — l/crhanum Auslraliacnsc. Canberra Figure 9 illustrates the fruit production of individually marked specimens of ten shrubs over ten years. The figure compares the production of each shrub with its production in the year when production was greatest. ■ aevok} sp^sce^i As-ac-a anejra •5. aoa prjincxofpo iy60 i96i 196‘^ Acacia lei'ogoncp'^} ''O •X'aoa dCO'ce .;s»o deioiQta I'assio he'm^i 1966 196/ i968 196.9 1970 I97| iremoprita >'Osef' Hakea 'area Sontckim spicatum O O 0 1-25 26-50 5' 75 '6-!00 Figure 9. — The crops of fruit produced by 10 species of trees and shrubs over JO years at Mileura .Station. Western Australia, expressed as percentages of the besl ycar\s production. Blanks indicate that no records were kept in those years (After Davies 1 976a). Three years stand out as ones when production was high, 1960. 1968 and 1971. Some fruit was produced in all years usually in the run-on sites. The fruits of shrubs thus provide a reliable crop for animals each spring and summer. In addition to these straightforward correlations, two modifying influences appear to operate on fruit production, one enviromenial and on biotic (Davies 1976a). In some species {Cassia desoUita and Acacia wiragonophylla) the fruiting of plants in some sites is correlated with winter rainfall but in others with summer rainfall. In these two cases the environment in the large creeks is so cold in winter that it inhibits the maturation of fruit set by summer rains: the only fruit that matures is that set in spring when winter rains arc heavy. Presumably the water stores in the soils of the large creeks are then suITicicnt for the plants to grow actively in the warm weather of the following spring, so that the fruit is matured. In the higher country the soils dry so quickly in the spring that soil moisture conditions may limit maturation after winter rain, but milder winter temperatures allow it to occur after summer rains. A second modifying influence is that of damage by animals, particularly insects and birds. In .1. adsurgens. A. aneura. A. letragonophylla. Eremophila frascri and E. freelingii animal attack has been seen to reduce or demolish a crop, often in seasons of high rainfall when a large crop would have been expected. Growth rates Germination and establishment of shrubs are rare events in the desert fringe. Information about the rates of growth of saplings was collected for only two species ,-i. adsurgens at the base of a granite monolith on Mileura and A. aneura near Wildlife Well in the Gibson Desert. The A. adsurgens grew steadily (Figure 10) over the years 1966-74 having probably germinated in 1961. Two of the five plants grew slowly but the other three showed rapid growth. The A. aneura (Figure 1 1) show'cd episodic growth, with one annual increment exceeding a metre and others of only a few centimetres (Table 1). The mulgas at the two sites near Wildlife Well showed 44 Journal oflhc Koval Society of Western Australia. Vol. 68. Part 2. 1986. Table 1 Mean heights and standard deviations of Acacia aneura shrubs at Wildlife Well (20° 50'S:125° 08’E) 1975-1980; rainfall figures for Wildlife Well are also shown. Year 1975 1976 1977 1978 1979 1980 Site 1 {n - 35) 115^43 180+79 187 + 94 204+101 226+108 Sile2(n= 15) 109±47 157+66 150+60 135 + 73 141±77 147+74 Rainfall (mm) Wildlife Well 119 + 9 0 76 0 208 different growth patterns. At site 1 growth was inhibited by lack of rain but at site 2, not only was it inhibited by lack of rain but browsing by camels Camelus dromedarhis resulted in negative growth in 1977 and 1978. The rain in 1978 was insufficient to stimulate a major pulse of growth but did enable the plants at site 2 to begin rehabilitation after browsing; the rain also provided the camels with more palatable food. The steady growth of .‘I. adsurgem appears to reflect the regular run-off it receives from the monolith; the episodic growth of the A. aneura reflects the rarity of a thoroughly wet season at a site in the central desert that receives little run-off. Animals Termites Davies (1 970) lists 24 species of termite recorded from Mileura, mostly collected by D. H. Perry and J. A. L. Watson. The resource partitioning of species of the genus Drepanotermes has been described by Watson Camels browsed the plants on sue 2 and caused the decline in ihc curve in 1 977 and 1 978. (1982). This genus of foraging termite has evolved species that differ from each other in preferred soil, preferred forage and methods of processing and storage of forage. Other species, especially those of Microcerotermes. Schcdorhiriolennes and Ileferoterme.s feed on decayed wood. Only two species build mounds Drepanotermes perniger and Tumulitermes tumuli, both foraging species. The mound of T. tumuli is frail and easily knocked from its base by passing stock so that most surviving mounds are close to and protected by thorny trees and shrubs. It is nevertheless still a common termite. The gross consumption of vegetable matter by termites on Mileura has not been measured but was estimated in New South Wales to exceed lOOkg/ha/yr. In a paddock near Alice Springs the biomass of termites was thought to be comparable with that of cattle (Watson, Lendon and Low 1973). Termites are important consumers of the plant material produced in the desert fringe. 45 Journal of ihc Royal Society of Wesiern Australia, Vol. 68. Part 2. 1986. Table 2 A list of grasshoppers recorded from Mileura Station, Cue* Gryllacrididae Hadrogn'llacris magnifica (Brunn.) Acrididae Alectoria superha Brunn. Aus!racris guttuhsa (walk.) Austroicetes anda Key Bermius sp. 2 Buforama sp. 2 Caperrala sp. 2 Chirotepica hisirio Sjosi. Chortoiceies lerrrnnifcra (Walk.) Coryphistes ruricula (Burm.) Desenaria hrtgirugosa Sjost Ecphatiius quadrilbus Stal FroggatUna ausiralis (Walk.) Goruaea ausirasiae (Leach) Happarana sp. 2 Happarana sp. 3 Macrolohalia ocellala (Tepper) Pycnosliciiis senauis Sauss. Qualetia maculata Sjost Stropis sp. 2 Stropis sp. Tapesta carncipes Sjost. Vrnisa guilulosa (Walk.) Vrmsiella sp. 1 Genus nov. 5 sp. 8 Genus nov. 6 sp. 2 Genus nov. 31 .sp. I Genus nov. 42 sp. 1 Genus nov. 72 sp. 5 (?) Genus nov. 83 viridis (Sjost) Genus nov. 91 sp. 2 Genus nov. 93 sp. I Genus nov. 95 ochracca (Sjost.) Genus nov. 95 sp. 4 *1 am grateful to W. Bailey and K. H. L. Key for assistance in the collection and naming of grasshoppers. Grasshoppers Davies (1970) lists twenty-two species of Acrididae collected on Mileura. Other species have since been collected and the known fauna is listed in Table 2. Between 1965 and 1971 seven I km transects were covered each September to obtain an indication of the numbers of grasshoppers available on the property. The catch was made by driving a vehicle at 40 km/hr steadily along a marked transect. .A grasshopper trap of the kind used by CSIRO (D. Clark pers. comm.), a steel mesh scoop with a catching face a metre square, was attached to the front of the vehicle. Figure 12 shows the total catch each year. The bulk of the 1966 and 1968 catches were Vrnisa gutiulasa. Most grasshoppers appeared to emerge after winter rain on Mileura and were at their peak of population in September. The figure indicates the great variation in numbers of grasshoppers present from year to year and therefore in the availability of food for animals that eat them. Torloises Only one species of tortoise Chelodina steindachneri lives on Mileura. It can be caught using meat as a bait in the permanent pools of Poonthoon and Ero, but in times of flood is found along the whole Pindabarn Creek and its main tributaries. Aestivating torloises have been unearthed from litter and beneath lawns round Murchison homesteads in summer. .After summer rains tiny torloises are sometimes caught in the creeks. From March 1960 to April 1961 Poonthoon Pool, where the Pindabarn turns west at the foot of the Jack Hills, was studied by setting 12 traps for two to three days each month in the pool. The results are presented in Table 3 in terms of the number of individuals captured, carapace width of individuals and dale of capture. One individual first caught in August 1960 was retrapped in November of 1965. The tortoises were difficult to trap in the winter when the water temperatures fell. Subsequent work in a tank in Perth indicated that an individual of this species did not feed when the air temperature was below 2 1 °C. 46 Journal orihc Royal Society of Western Australia. Vol. 68. I^art 2. 1 986 Table 3 Results of trapping for the tortoise Chelodina steindachneri in Poonthoon Pool. March 1 960 to April 1961. Carapace Width (cm) Month MAMJ J ASONDJ FMA 7-10 1 10-13 3 11 2 1 13-16 2 2 ' 1 16-19 1 Totals 00031*421*0*31 * — no trapping undertaken Zebra Finches From 1973 to 1981 a population of Zebra Finches Poephila guttata was studied at Jindi Jindi mill on Mileura Station, a mill that lay about 7 km from the creek in a section of the Belele land system where much wind grass Aristida contorta grows. The seeds of this plant arc an important food of the Zebra Finch at Mileura (Davies 1977al. Early studies showed that the birds bred in both spring and autumn but those bred in the spring survived best. Accordingly, in order to sample the population an attempt was made to catch a large sample each year in October from 1974 to 1981. so that the population could be estimated by a capiure/recapturc method. Trapping lasted several' days and the population was estimated each da\ after the first on the basis of that day's recaptures. In 1981 1 398 finches were trapped and marked over four days. During the next two days a total of 600 were trapped at the six adjacent water points but no birds banded at Jindi Jindi were recovered. U was concluded that emigration and imniigraiion from Jindi Jindi during the short capture period each year could be ignored. The population estimates and proportion of juveniles in the catches are shown in Table 4. Table 4 Estimates of the size of the Zebra Finch population and the proportion of juveniles in each >ear's sample at Jindi Jindi mill. Mileura Station. W. A. Date No. days sampled Estimated population Percentage of juveniles Total trapped Total retrapped 1974 2 14 700 247 4 1975 2 3 748 1 1 250 13 1976 3 686 0 137 1 1 1977 4 810 0 85 9 1978 4 2 949 21 244 10 1979 3 1 478 6 155 10 1980 2 22 258 35 351 5 1981 4 18 578 25 1 398 58 Of the total banded at Jindi Jindi over those years (2 867) 120 were rclrapped. Only two of these were away from Jindi Jindi (7 km) and of the others one was 3 years after banding, five 2 years after banding, 19 one year after banding and the remainder (95) in the year of banding. Few Zebra Finches probably live more than three years in the wild at Mileura. The distribution of Zebra Finches at Mileura followed that of the grass plains. They were most abundant at water points near those plains and scarcer in the Sherwood and Mileura/Berringarra land systems, the hills and the creeks respectively. The large fluctuations of the Zebra Finch population at Jmdi Jindi resembles that of the grasshoppers and suggests that both breed prolifically when conditions are good; most die but a few survive over the dr>' years. Autumn breeding is presumably an advantage in a species that can breed at three rnonlhs of age because, when numbers are very low. the survival of even a few will boost the potential breeding population in the spring (Davies 1979). Emus The Emu Dromaius novaeholfandiae is a herbivore of the arid inland, often living in great numbers on Mileura. Studies of various aspects of the bird’s biology have been made (Algar 1980, .Ambrose 1980. Beutel et al. 1983. Davies 1968. 1976b, 1977b. 1978. 1984, Davies el al. 1971. Davies and Curr>' 1978). In summary Emus live as pairs from December to May. maintaining a home range of about 30 sq km in which the female lays a clutch of 9-20 eggs from April to June. The size of the clutch varies with the amount of summer rainfall. Once the clutch is laid the male takes over, incubating the eggs for eight weeks without eating or drinking. When the chicks hatch he leads them for five to seven months after which they become independant (Curry 1979). The male then remates and starts a new breeding cycle, probably rarely with the same mate. Emus feed on fruits, seeds, flowers, insects and the young growing parts of plants. Items satisfying this diet occur in abundance ai particular sites only irregularly, so that the Emus need to be continually able to move to maintain contact with their foods. It is only when the male is incubating that he cannot move. He does not eat, drink or defecate then, so that his immobility is not detrimental. Al all other times Emus have no ties that prevent movement and arc completely nomadic. The Emu's biology is particularly well adapted to life in the desert fringe where food and water arc erratically distributed season by season. Movements of banded birds show that individuals may remain in one spot for over a year or move many hundreds of kilometres in a few months, presumably maintaining contact with a good food supply or attempting to do so (Davies 1984). The desert fringe at Mileura produces a reliable crop of seeds and fruit each spring and summer (Davies 1976a) but the autumn and w inter crop of annuals is sometimes sparse. Emus sometimes move south-west out of the desert fringe in large numbers in winter, apparently in response to a shortage of food, orientating their movements towards recent rainfall events (Davies 1984). These movements are of recent origin; the first reported in 1906 (Rogers 1906). The establishment of additional watering points by the spread of pastoral development has enabled Emus to live permanently away from the large pools of Cycle 3 creeks (Figure 5). In the past they could live on Cycle 1 and 2 creeks only after rain until the small natural pools dried up. as Emus appear to need to drink each day. even in cool weather (Davies 1972). Therefore in favourable seasons many more Emus can now live and breed in the desert fringe than it can support in dry seasons. The bird’s behaviour leads them to leave places where they meet many other Emus. As food becomes sparser their search for it becomes more extensive, increasing the chance of meeting many other Emus. So as the small pastures of the Cycle I and 2 creeks, that could support birds after good rains, become exhausted the Emus from those places move into the pastures of Cycle 3 creeks. Before pastoral development few birds were involved but now there are too many to be supported in the more favoured 47 Journal ofthe Royal Society of Western Australia. Vol. 68. Part 2. 1986 sites and a large scale exodus is gradually initialed leading to such spectacular movements as took place in 1959, 1969 and 1976 (Davies 1983). Such movements lead to the death of many birds. Few Emus are found in land not yet developed for pastoral use, supporting the suggestion that such development may have led to an increase in the numbers of Emus (Davies 1969). Emus represent the nomadic answer to survival in the desert fringe. They arc adapted to maintaining continual contact with their food supply, suffering catastrophic mortality in bad seasons and producing many offspring when conditions are favourable. Lillie Brown Bats From 1965 until 1982 a small colony of Little Brown Bats Epetesicus pumilis was studied in Ejah and Flemington paddocks at Miieura. Over that time 356 were marked. The largest catch was of 94 and the smallest three. Of the 60 recaptured at least once 28 were caught after one year, 16 after two, 1 1 after three, two after four and 3 five years after banding, indicating that this liny animal weighing less than 5 g. may survive at least five years in the wild. The colony is probably site specific. Of three marked bats taken 40 km away and released, two were later recaptured at the colony. The species probably breeds well and builds its numbers up in good seasons, a few surviving during poor ones. The best season for the bats, 1976, was one of the worst for many other species. Pebble-mound Mice The little Pebble-mound Mouse Pseiidomys chapmanii once lived on Milcura. It does so no longer but has been captured on the Hammersley Range. It characteristically builds a pebble mound, each pebble within a weight range of 1.5-3.8g and a volume 0.6-1. 7 cc. These mounds can be 2 m across and 1 m deep, the uniformity of their pebbles such that they are used as screenings for concrete making by pasloralisls. No one is certain how the mounds arc used but it is surmised that they act as dew ponds, cooling at night and heating more slowly than the surrounding air as the sun rises, so that water condenses onto them. In this w'ay the mice could obtain water in places where the soil is too shallow to allow them to dig deeply to moist soil, usually about 60 cm below the surface. In 1976 it W'as possible to survey the distribution of these pebble mounds in the Murchison and Pilbara. Figure ! 3 contains data from Dunlop and Pound (1981) and data from Davies, Knight and Rooke (pers. comm.) showing the mounds to be widely distributed. At one site we mapped the mounds. Figure 14, finding to our surprise that there were, intermingled with the small pebble mounds, mounds of large pebbles (weight range 1 5.5-26 g: volume range 6.5 to \ 2.0 cc). The distribution of the large pebble mounds differs from that of the small and is more southerly. Burbidge (pers. comm.) has suggested that they may be relicts of the nests of a slick- nest rat, rather than a giant dew pond mouse and I am inclined to agree. The Pebble-mound Mouse has left behind a series of public works that must have taken years to build and been used by generations of mice. Perhaps all those at one site were not occupied at the same time. Discussion Coming to terms with the environment I. Spatially The Tortoise, the Bat. the Pebble-mound Mouse and the Echidna have come to terms with the environment of the desert fringe by using nests placed in sites that, in one way or another protect them from the environmental rigours of the area. They come out to feed w^hen conditions arc to their liking. At other times of day or season they retreat and hibernate or aestivate. Olhei species exploit the creek systems, using these favoured sites where water accumulates as the base from which to penetrate the 90% of the land surface that is unfavourable. The scale varies. Zebra Finches use the small watercourses where the annual grasses grow. Emus the large creeks that support shrubs, but the principle is the same. Both depend on only part of the landscape although they seem, at first sight, to be everywhere. 48 Journal of ihc Royal Sociely of Wesiern Australia. Vol. 68. Pan 2. 1 986. Figure 14. — ,A map of a site on F.rong Station. Western Australia, showing the positions of pebble mounds surmised to have been built by Pseudotnys chapmani (small dots) and an unknown larger Murid (large dots) in a mi.xed colony. 2. Temporally Some animals live on the fringe of the desert by moving to keep in touch with a food supply. The Emu is one. Its nomadic behaviour differs only in degree from that of a resident territorial bird that uses different parts of its territory at different limes of the year because the resources it needs arc at dilTerent places at different limes. Somehow Emus can detect where resources are from great distances. This ability poses a fascinating unsolved problem of the desert fringe, how do they do it? The desert fringe provides a mosaic of living conditions for animals, some ven stressful, some no more extreme than those of the woodlands near the coast. Only highly adapted desert forms can live in the most exposed and arid sites; they flourish when conditions are dry but arc inconspicuous in wet years. The redistribution of water means that in some places regular seasonal flushes of growth can occur. In such places animals requiring regular resources can persist, expanding to other areas when rainfall is heavy. Provided an animal can maintain a population, no matter how small, it can become part of the desert community, abundant only when conditions favour it. often existing as sparse, disjunct populations. Each species requires a different degree of wetness: not all are favoured by very heavy rains, so that in each year some organisms find conditions especially favourable. So the dominant animals and plants change year by year, a seemingly endless variation drawn from a common pool. This effect is easy to illustrate on the desert fringe, yet it is true too of more mcsic environments. It is easy to overlook the irregularities of climate and the consequent changes of dominance in woodlands, forests and jungles. Studies of the desert fringe sharpens one's perception of the dynamics of all environments. Perhaps the most interesting conclusion to be drawn from this study is that many successful, that is surviving, organisms of the desert fringe, live more than one year. Even small ones like the bat and the Zebra Finch have been shown to sur\dve 3-5 years. Considering the rainfall pattern that show's 80-90% reliability of enough rain falling to flood the watercourses, this makes sense. The organisms can last for several years without breeding but then when opportunity offers, breed rapidly and repopulate. as the figures have shown the grasshoppers, finches and bats can do. The annual plants seem to be the organisms that best give the lie to the belief that everything is erratic on the desert fringe. Their seed is viable for only a year. Reproduction must be regular, only production is erratic. Other organisms need more resources to breed at all, but even for them a 3-5 year life span seems to enable them to pass the test of survival in the desert fringe. Acknowledgements . — The work al Mileura owes much lo the unfailing hospitality and help we have always received from Mr. T, F. M. Walsh and his family, owners of the station. Many other station owners and residents of the pastoral areas of Western Australia have helped me in many dilTerent ways over the years that 1 have spent working in the Murchison and adjacent areas. CSIRO supported the work generously and the liniversily of W'estem .Australia. Murdoch University and the Western .Australian Institute of Technology have all contributed in various ways. My Ikld companions. M. W. R. Beck. K. Casperson. H de Rebeira, T. .A. Knight. J I*. Kruiskamp. J. J. Moil. T. C. Scott, and D. Scott- Smilh have helped me through many years of data gathering, 1 am grateful for their patience and skill, and also to P dc Rebeira, F. knight and B. (^uin who prepared the figures and A. R. Stlbersietn who read the text critically for me. References Algar, D- (1980). — Studies on growth, body composition and nutrition of Emus. MSc. Thesis, University of Melbourne: RAOU Microfiche Series No. 9. Ambrose, S. J. ( 1 980). — Comparative study of the water metabolism of the Emu. Galah and Zebra Finch. Unpuhl. Hons. Thesis. University of Western Australia. Arnold. J M. (1963). — The climate of the Wiluna-Meekatharra area. CSIRO Land Res. Ser . No. 7; 73*92. Barker. W. R. and Grccnsladc. P. J. M. (1982). — Evolunon of ihe flora and fauna of and Ausiraha. Adelaide. Peacock. Beutel, P M-. S. J. J. F. Davies and W. C. Packer. (1983). — Physical and phv-stological measurements of Emu eggs and chicks. Iniern. Zoo Hundl>o<>k,2i: 175-181. Curry, P J.(I979). — The young Emu and its family life in captivity. MSc. Thesis. University of Melbourne; R.AOU Microfiche Series No. 1. Davies, S. J. J. F. (1968). — Aspects of a study of Emus in semi-arid Western Australia. Prnc. Ecol. Soc AuM . 3: 160-166. Davies. S. J. J. F. (1969), — An aerial survey of inland W'estern Australia. Tech Memo /)/»■ Wdd. Res. CSIRO. Aust. No. 1. Davies, S. J. J. F. (1970). — A list of plants and animals found at Mileura, Western Australia. Lat. 26‘‘22'S. Ixing. 1 17° 20'E. Icch. Memo. On Wild. Res. CSIRO. Ami. No. 3. Davies. S. J. J. F. (1972). — Results of 40 hours' continuous watch at five waterpoints in an Australian desert. EVui/, 72: 8-12. Davies. S. J. J F. (1973). — Environmental variables and the biology of native Australian animals in the mulga zone. Trap. Grasslands. 7: J 27- 134. Davies. S. J. J, F. (1975). — Land use by Emus and other wildlife species in the arid shrublands of Western Australia. Proc. 3rd. Workshop CS/.Ausi. Rangelands Panel. Tuscan. Arizona. 1973: 91-98. Davies, S. J. J. F. (1976a). — Studies of the flowering season and fruit production of some arid zone shrubs and trees in Western Australia. J. Ecol. 64: 665-687. Davies. S. J. J. F. (1976b). — The natural history of the Emu in comparison with that of other rallies, Proc. }bih. Intern. Ornith. Cong.: 109- 120 . Davies. S. J. J. F. ( 1976c). — Environmental variables and the biology of birds of the Australian arid zone. Proc. I6rh. Iniern. Ornith. Con,?.: 481-488. Davies, S. J. J. F. (1977a). — The timing of breeding by the Zebra Finch Taeniopvgia castonotis at Mileura, Western Australia. Ibis. 119- 369-372'. 49 Journal of the Royal Society of W estern Australia. Vol. 68. Part 2, 1986. Davies, S. J. J. F. (1977b). — Man’s activities and birds’ distribution in the arid zone. Emu. 77: 169-172. Davies, S. J. J. F. (1978). — The fcK>d of Emus. .-Iwj-/. J. ecoi, 3: 41 1-422. Davies, S. J. J. F. (1979).— The breeding seasons of birds in south-western .Australia. J Roy. Soc. H .-i., 62: 53-64. Davies, S. J. J. F. (1982). — Behavioural adaptations of birds to environments where evaporation is high and water is in short supply. Comp. Biochem Physiol.. 71A: 557-566. Davies, S. J. J. F. (1983).— Emus in Western Australia. In Riney, T. (Ed.) Pi'ildlife Management in the 80's. Melbourne: Graduate School of Environmental Science. Monash University: 240-247. Davies, S. J. J. F. (1984). — Nomadism as a response to desert conditions in Australia. J. .Arid Environ..!: 1 83-1 93. Davies. S. J. J. F.. Beck. M. W. R. and Kruiskamp. J. P. (1971). — The results of banding 154 Emus in Western Australia. CSIRO Wildl. Res.. 16; 77-79 Davies, S. J. J. F. and Curry. P. J. (1978). — Emus. Proc. No. 36 Course for Veterinarians. Feb. 6-10, 1978. Univ. Sydney; 31 1-317. Davies, S. J. J. F., Knight, T. A., Pfiuner. G. H. and de Rebeira, C. P. S. (1977). — Ejah Breakaways. Mileura. Cue, as an aboriginal home. If. Aust. Nat.. 13: 177-183. Davies, S. J. J. F. and Walsh, T. F. M. (1979). — Observations on the regeneration of shrubs and woody forbs over a ten year period in grazed quadrats on Mileura Station, Western Australia, .iust. Rangel J . 1: 215-224. Dawson, W. R. (1976). — Physiological and behavioural adjustments of birds to heat and aridity. Proc. !6th. Intern. Ornilh. Congr.: 455- 467. Dunlop, J. N. and Pound. I. R. (1981). — Observations on the Pebble- mound Mouse Pscudomys chapmani Kitchener. 1980. Rec. West. AiLSt. Mus.. 9: 1-5. Fnih. H. J. (Ed.) (1976). — Proceeding of I6ih. International Ornithological Congress, Canberra. 1974. Canberra. Australian Academy of Science. Griffiths, M. (1968). — Echidnas. London. Pergamon. Hamngton, G. N.. Wilson. A. D. and Young. M. D. (1984). — Management of Australia's rangelands. Melbourne, CSIRO. Keasi. J. A. (Ed.) (1981). — Ecological Biogeography of Australia. The Hague, Junk. Mabbutt, J. A. (1963). — Introduction and summary description of the Wiluna-Meekatharra area. CSIRO Land Res. Sen. No. 7: 9-23. Mabbutt. J. A.. Litchfield, W. H.. Speck, N. H,, Sofoulis. J., Wilcox, D. G.. Arnold, J. M.. Brookfield. M. and Wright, R. L. (1963). — General report on the lands of the Wiluna- Meekatharra area. Western Australia. 1958. CSIRO Land Res. Ser. No. 7. Mott. J. J. (1972a). — The autecology of annuals in an arid region of Western Australia. PhD Thesis. University of Western Australia. Mott. J. J. (1972b). — Germination studies on some annual species from an arid region of Western Australia. J. Ecoi. 60: 293-304. Mott, J. J. ( 1973). — Temporal and spatial distribution of an annua) flora in an arid region of Western Australia. Trop. Orasslands, 7: 89-97. Mott. J. J. (1979). — Flowering, seed formation and dispersal. In Perry. R. A. and Goodall. D. W. (Eds). Arid Land Ecosystems: Structure. Function and .Management. Vol. 1. IBP 16: 627-645. Nix. H. A. (1982). — Environmental determinants of biogcography and evolution in Terra Australis. In Barker, W. R. and Greenslade, P. J. M. (Eds). Evolution of the flora and fauna of arid Australia. Adelaide. Peacock: 47-66. Preece, P. B. (1971).— Contributions to the biology of mulga. 11. Germination. / Bot.. 19: 39-49. Rogers.;. R. (1906).— Do Emus migrate? £m«. 6: 13-14. Serventy. D. L. ( 1 97 1 ). — Biology of desen birds. In Famer. D. S. and King, J- R. (Eds). Avian Biology^ New York. .Academic Press: 287-339. Slayler, R. 0. ( 1 96 1 ). — Methodology of a water balance study conducted on a desert woodland {.Acacia aneura F. Muell.) community in central Australia. Arid Zone Res.. 16: 1 5-25. Watson. J. A. L. (1982). — Distribution, biology and speciaiion in the Australian harvester termites. Drepanolermes (Isoplera; Termitinae). In Barker, W. R. and Greenslade, P. J. M. (Eds.) Evolution of the flora and fauna of and Australia. Adelaide. Peacock: 263-265. Watson. J. A. L., Lendon, C and I-ow, B. S. (1973), — ^Termites in mulga lands. Trop. Grasslands, 7: 121-126. Watson, J. A. L. and Perry, D. H. (1981). — The Australian harvester termites of the genus Drepanolermes (Isoptera: Termitinae). Aust. J. Zool. Suppl. Ser., 78: 1-153. 50 Journal of the Royal Society of Western Australia, Vol. 68, Part 3, 1986. Terminology for geomorphic units and habitats along the tropical coast of Western Australia by V. Semeniuk 2 1 Glenmcre Road, Warwick W.A. 6024, .Australia Manuscript submitted: '^(h September. 19H4: accepted 15th October 1985 Abstract The tropical coast of Western Australia comprises a large range of shoreline types and coastal features that require categorisation, and this paper provides a nomenclature system to describe the geomorphic units and habitats of this region. A review of inlernaliona! and local literature concludes firstly that, at the regional scale, classifications tend to be too genetic to be of use in studies of the tropical Western Australian coastline, and secondly, that few studies have come to terms with either a nomenclature or philosophy of approach that deals with the various scales of coastal features. This paper provides an approach to describing coastal features by utilising a nominated, fixed scale as a framework to nomenclature, and also provides a terminology by defining terms for the various scales of coastal features. The frames of reference in decreasing scale are defined as: regional scale, large scale, medium scale, small scale and fine scale. Within each frame of reference there are a variety of geomorphic units that arc distinguished on criteria of depositional/erosional setting, geometry, morphology of surface, substrate, land surface position and geomorphic processes at surface. Coastal landforms thus can be systematically described in progressively decreasing scales, and a coastal type may be classified as a geomorphic unit at a fixed, nominated scale. The regional to fine scales of geomorphic units also may be used as a framework to distinguish types of habitats for organisms along a coast. Introduction Coastal environments have been studied by numerous authors in a wide range of scientific disciplines. As a result there is much literature dealing with classification, nomenclature, processes, products and principles appropriate to the scale and type of study be U biologic, sedimentologic, geomorphic, etc. Summaries of such studies are presented in lexis by Cotton ( 1 952), Valentin (1952), Davies (1964, 1980), Holmes (1965). King (1972), Bird 0976a), Chapman (1976), Bloom (1965, 1978), and Davis (1978). Other examples of specific studies in various disciplines of geology, sedimentology and biology are covered in Dyer (1973), Ginsberg (1975), Bird (1976b), Chapman (1976). Langford-Smith and Thom (1969), Day (1981). Stephenson and Stephenson (1972), and Wolff (1983). In general, because studies are specifically oriented towards (1) a given discipline, or (2) a particular scale of reference, or (3) a classification objective, there has developed a diverse range of classification and nomenclatural systems which are only partly applicable or useful to all aspects of coastal science. For example, the classification of regional tectonic and morphologic coastal features by Inman and Nordstrom ( 1 97 1 ) is at an inappropriate scale and emphasises factors largely irrelevant to the biologist who requires a classification of small to medium scale features, a scale at which biota and habitats develop and interact. Conversely, the small- scale differentiation of sediment units or habitat units as described in Cooper (1958), Ginsberg (1975), and Goldsmith et al. (1977) is inappropriate (i.e. far too detailed) fora study of regional classification as required by Jennings and Bird ( 1 967). The northwest coastline of tropical Western Australia, north of Pt Cloates through to Cambridge Gulf, comprises a large range of shoreline types and coastal features that require categorisation so that a consistent multidisciplinary terminology can be applied. In order to avoid problems that have developed with other classification systems, it is proposed here that a more rational conceptual framework and terminology be adopted in studies of coastal geomorphology and habitats in tropical W'estern Australia as a prelude to further work on geomorphology, sedimentology, stratigraphy, hydrology, oceanography, chemistry and biology in this region. The need for codification of terminology is necessary because of the amount of research on geomorphic units and habitats intended along the Western Australian coast as a whole, and because there already has been an inconsistent use of terms and concepts applied to geomorphic/habitat units. The aim of this paper therefore is to provide a nomenclature system to describe geomorphic units and habitats of the tropical coast of Western Australia. However, prior to developing a nomenclature system, a review of global and local literature is presented so that the precedence of other workers can be assessed. In detail, the paper thus provides: (1) a review of international and local literature on coastal geomorphology, (2) an approach to describing coastal features for tropical Western Australia utilising a nominated scale as a framework to nomenclature, and 46670-1 53 Journal of ihc Royal Society of Western Australia. Vol. 68. Part 3, 1 986. Figure 1. Location of study area between Exmouth Cult and Cambridge Gulf. The lower part of the figure illustrates the scalar frames ol reference. In each case here the lower limit of scale is used in the frame. 54 Journal of the Royal Society of Western Australia. Vol. 68. Part 3. 1986. (3) a terminology, by defining terms for the various scales of coastal features. Much of the philosophy and description presented in this paper is scattered in various texts and scientific journals. However, in most cases the published works only deal with aspects of the approach presented here, or present the end products of classification (i.e. terminology) rather than a philosophy of approach to gcomorphology which can be applied explicitly to more than one discipline. The term coast as used here is intended to encompass the shoreline interface between land and sea as well as those features immediately landward of the shore. The term ‘coastline’ thus encompasses the tidal zone and the adjacent subaerial (supratidal) strip. Methods The primary method used to obtain information for this paper has been fieldwork. Some 12 months over the past 10 years have been spent in the field in a wide variety of coastal settings along the north-west and north coast of Western Australia (Fig. 1). Fieldwork has involved: mapping terrain and coastal features onto aerial photographs; documenting geometry of terrain/habiiat units, including their surfaces, substrates and interfaces: and collecting subslrale/soil and water samples. The remainder of the coast was surveyed and photographed during low-levcl flight by light plane and helicopter. Fieldwork was supplemented by examining the aerial photographs of coastal sections not amenable to access. Global Review Numerous classification schemes and their accompanying terminology have been established w'orldwidc for natural coastal units. It is worthwhile to review some of these as a basis for precedence in either nomenclature or philosophy of approach for the Western .Australian coastline. (icneral C 'fassifications Regional classifications presented by authors in coastal gcomorphology tend to be genetic, at least at the higher levels of heirarchial organisation (see Johnson 1919: Cotton 1942. 1952: Valentin 1952: Price 1955; Shepard 1963: Davies 1964; Bloom 1965). While this approach is appropriate to understanding the origins of coasts and assessing different factors that lead to coastal variabilitv, it is not altogether useful for coastal workers w'ho require a descriptive framew'ork for their studies. Furthermore, much of these classification systems are concerned with regional scale or large scale features. They do not deal with the smaller scale divisions necessary for biologists, sedimenlologisis and (process- oriented) geomorphologisls. Finally. and most importantly, classification must build on a foundation of descriptive studies and not vice versa (see RussclM967). There are numerous instances of genetic classification systems that do not predict and therefore do not allow for some specific coastal categories. In practical terms, classifications should be constructed with the hindsight of available information. For these reasons the classifications described m the works cued above are considered inappropriate to this study. Smaller-scale Subdivisions Even though the established genetic classification of coastal landforms is rejected here as a basis for catergorising the coastline of tropical north-west and north Australia, the philosophy of approach that has evolved for smaller-scale coastal subdivision, the criteria for smaller-scale subdivision and the terminology that have been developed worldwide have some applicability. How'cvcr many of the criteria of classification and the resultant terminology are specifically oriented towards particular coastal systems and are not applicable universally; for example, the classification criteria and terminology of units within a delta obviously arc not applicable to units of barrier dunes. The principle of subdivision into units also is utilised in many different coastal settings and in different disciplines. The work of Zenkovilch (1967) is a useful example of this principle. Zcnkovitch (1967) provides a descriptive approach, as well as a suite of terms, in classifying various types of sedimentary deposits along steep indented shorelines. For the classification Zenkovilch { 1 967) utilised primar>' non-gentic criteria of morphology (slope, orientation, position, secondary shape features, and quantification of some parameters), but also utilised dynamics and genetics. The classification incorporates many scales of reference; it allows description of coastal forms in detail, and also provides information and insight into processes of evolution and maintenance. The approach of Zenkovilch (1967) is a good example of coastal description and classification that should be emulated. Similar internal classification of specific coastal landforms has been accomplished in delta areas, coastal dune systems and barrier island/proiecled tidal fiat systems (Cooper 1958. Allen 1970. Coleman ei al. 1970, Gould 1970. Purser and Evans 1973, Coleman and Wright 1975. Evans e{ al. 1977, Goldsmith e! al. 1977, Goldsmith 1978. McKee 1979). In delta areas for instance, depending on the need for detailed subdivision, authors have identified finer scale units using various criteria relevant for iheir explicit purpose: the Mississippi della has been subdivided into numerous gcomorphic/sedimcniologic units as a framework to scdimeniologic-siraligraphic studies (Fisk et al. 1954, Fisk 1961. Frazier 1967, Ciould 1970): the Tabasco delta has been subdivided into geomorphic-stratigraphic units as a framework to biological studies (Thom 1967); barrier island coasts have been subdivided into medium- scale units for the purposes of stratigraphic and biological studies (Hayes 1975. Phclgcr 1977) and small- scale units for purposes of sedimcniologic studies (Hayes and Kana 1976). On the other hand geomorphologists in these areas have tended to recognise units for mapping purposes using mixed criteria, attempting to provide a geomorphic framework al various scales for studies such as surface processes, soils, vegetation and land use. The .scale at which a study is organised is determined by the type of detail required. Obviously detailed substrate morphology, which is of relevance to biologist or a scdimentologisl' is largely irrelevant to a regional coastal geomorphologisi who only needs to identify large-scale components. Conversely the scale at which a siudv terminates depends on whether there is a need for finor-scalc information. The scdimentologisl and biologist both can utilise geomorphic information at large scale (c.g. deltaic setting; cf. Wright 1978). medium scale (e.g. beaches/dunes within the deltaic selling) and small scale (c.g. dune crest, dune swale, back shore of the 46670-2 55 Journal of ihc Royal Society of Western Australia, Vol. 68. Part 3, 1986. beach, and dune setting), but they also can utilise information at progressively smaller .scales, and studies such as these then fall out of the realm of traditional geomorphology and into sedinientology sensu stricto (e.g. in further decreasing scale there are: large-scale bedform surfaces small-scale bedform suiiaces. variability of grain size, variability of grain types). On a worldwide basis, terminology appears to have followed the pattern listed below, even though the pattern has occurred fortuitously: (1) Recognition of systems at the regional scale (the philosophy of this approach is covered by Mitchell (1973) and Bloom (1978) for terrestrial as well as coastal systems); usually a primary criterion utilised is whether the coastal units are depositional/erosional or submerged/emergent (Johnson 1919. Valentin 1962. Shepard 1963). (2) Identification of coastal landforms (sediment bodies or erosional interfaces) based on their geomeir\. their phoioione on aerial photographs, substrate type and biota (such as macrophytes). (3) Further subdivision of the coastal system based on substrate differences, small-scale geometry, tidal levels and biota. This pattern is one where the development of terminology and classification has been inadvertently scale-determined. Discussion The enormous wealth of terms that has been coined in geomorphicAsedimcniologic studies of coastal areas is not altogether satisfactory for use in Western Australia. Certainly there arc a sufficient numbci of relevant terms for large-scale gconiorphic features such as deltas and ria coasts, but many terms for smaller scale features are non-existent m the literature or not applicable to the tropica! coast of Western .Australia. For instance, the classic, established system of gcomorphic subdivision of a tidal flat (van Straalen 1954) envisages a high tidal flat, intertidal slope and subtidal zone. These units or their conceptual equivalent have been utilised subsequently by Thompson (1968). .Allen (1970). and various workers in Ginsberg (1975). Howe\er, the units are not strictly relevant to the tidal coastline of nurih-wcsi and north Australia. Similarly, terms to describe units peripheral to limestone barrier islands, rocky shores, ria shorelines arc also inadequate. The literature, however, has provided useful terms for the following coastal settings, particularly at large and medium scales hut less so for the small scale: (!) deltas. (2) beach/dunc coastlines. (3) sandy barrier islands, and (4) spit.s/chenicrs and tombolos (Frazier 1967, Morgan 1967. Zenkovitch 1967, Allen 1970, Coleman ef ul. 1975. Haves 1975 Hayes and Rana 1976, Phleger 1977. Boothroyd 1978 Davis 1978. Wright 1978). The main conclusion of this global literature review on coastal terminology is that, as scale of reference decreases and numbers of gcomorphic units/cnlilics increase, the terminology becomes less adequate or relevant to the study area of this paper. This is not sun>rising. since much of the tropical Western Australian coast is globally unique and it is to be expected that there mav be undescribed combinations of gcomorphic processes and landforms. Where globally established terminology is adequate or relevant to the north-west and north Australian coastline it is utilised later on in this paper. Review of Literature on Central VVest/North-west/North Coastline of Western Australia Introduction A number of papers have already described segments of the central west, north-west and north Western Australian coastal and near-coastal (shallow-water marine) environments. Although not all of the studies are located in the tropical zone, it is worthwhile to review the main works here in order to appreciate what precedents in approach and nomenclature have been set. Specifically Jutson (1950), Fairbridgc (1951). Russel! and McIntyre (1966), Jennings and Bird (1967), Logan and Ccbulski (1970). Jennings and Coventry (1973). Wright et af. (1973). Brow'n and W'oods (1974). Hagen and Logan (1974). Read (1974). Jennings ( 1 975). Thorn et al. (1975). Woods and Brown (1975). Logan and Brown (1976), Davies (1977), Geological Survev of Western Australia (1980. 1982a. 1982b. 19^2c) Semeniiik (1980. 1981a. 1981b. 1982. 1983. 1985), Galloway (1982). Johnson ( 1982), Senieniuk el ul. (1982) and Hesp and Craig (1983) have published studies on geomorphology and sedinientology of the tropical Western Australian coast. In general the terminology, classifications and philosophy of these works follows that outlined in the Global Review. (nmeral Studies Many of the .studies that deal with aspects of coastal geomorphologN arc concerned with regional scale aspects and so pro\ ide only a regional selling and listing of large-.scale components (c.g. Jutson 1950. Fairbridge 1951. Jennings and Bird 1967. (Jcological Survev of Western Australia 1974, Davis 1977, Gallowav 19'82). Jennings and Bird (1967) for example, identi'fy King Sound as a regional gcomorphic unit, terming it an e.stuary. and do not proceed beyond identifying alluvial plains, tidal mud flats, mangroves and shoals. Publications by the Geological Survey of Western Australia (198(), 1982a. 1982b, 1982c) similarly onl\ generally identify broad components of the shoreline and coast (e.g. mud flats, sand dunes, limestone rcefs/cliffs/ouicrops. etc.), (iallowav (1982) provides a broad description of the coastal lands of Nonh-Wesicm Australia as part ot a regional description of ph>;siographic patterns associated with mangroves. Wright et at. (1973) similarly categorise the north-west coast of Australia into distinct provinces between the east Kimberley and Danvin. Davies (1977) provides a concise chapter on the whole Australian coastline and identifies regional components such as rocky coasts, tidal plain coasts, barrier island coasts, and relates these to the major influences of geological structure and large- scale processes. While relevant to an understanding of the main factors that develop different coastal types at regional and larger scales, these approaches of Davies (1977) and VVrighl ei ai (1973) are largely inapplicable to studies at a more detailed level, or to studies that require a descriptive framework. On the other hand some studies are onlv rcconnaisance. Russell and McIntyre (1966) in a brief Auslralia-vvidc study describe a variety of tidal flats in tropical Western Australia. Although the various tidal zones are not allocated precise terms, the local study areas of these authors were described in some detail along selected transects. Hesp and Craig (1983) mention coastal landforms in a study of Pilbara coastal flora but provide a very incomplete picture of coastal 56 Journal of the Royal Society of Western Australia. Vol. 68. Part 3. 1986. gcomorphology. Out of some 9-10 large- to medium- scale units obvious along the Pilbara coast, Hesp and Craig describe only three inter-related units and provide sketchy mention of the others. Specific Studies The remaining papers generally concentrate on coastal evolution, coastal sedinicntology or marine habitats, and utilise geomorphic units as a framework to the specific studies. All these studies, however, are relevant to the philosophy of this paper because they employ geomorphic terms that extend from regional through to medium/smail scale obscr\'ations. These papers arc reviewed below in terms of: the approach used by the author/s. the criteria utilised for subdivision of coastal units, and the terms used to name the subdivisions. In a scries of papers Logan and colleagues {op. cit.) describe the Shark Bay coastal and marine system, primarily from the point of view of carbonate scdimcntology and evolution of stratigraphy. In general they have followed the global precedent: large-scale units were identified and later subdivided into smaller units as the studies required. The basic paper by Logan and Cebuiski (1970) describes the large-scale geomorphic system as a framework for the sedimenl- ology/slraligraphy of Shark Bay as: (1) Embaymcnl Plains and Basins; (2) Sublittoral Platforms; (3) Sills: and (4) Intcnidal-supraiidal zone. There is further subdivision of these large-scale features into finer scale units based on substrate dilTcrenccs, tidal levels and slope (c.g. intcrtidaLsiipralidal zones are subdivided into rocky intertidal areas, intertidal beach areas and lidal- supratida! flats). Subsequent authors, e.g. Brown and Woods (1974) and Read (1974). have adopted the terminology /classification of Logan and Cebuiski (1970). but modified and subdivided the units when necessary. Read (1974). working on scagrass platforms and sills, identifies smaller scale geomorphic entities of tidal channels, mcgaripples and sand ribbons as subsidiary elements of sills; Brown and Woods (1974). Hagan and Logan (1974) and Woods and Brown (1974) working on selected tidal flats of the region, subdivide the lidal- supralidal zone into six units based on substrates and levels above low tidal datum using terms such as beach ridges, supratidal flat and high intertidal flat. Johnson ( 1 982) in a scdimenlology/stratigraphic study of the Gascovne delta subdivided the deltaic system into a series of medium-scale geomorphic units termed bar unit, bank unit, strand plain unit (composed ot beach ridges and tidal flats), channel unit, levee unit and flood plain unit. Smaller-scale geomorphic units within these geomorphic entities were noted in the description but not speciflealiv nominated because the study endeavoured only to identify medium-scale units as a basis for stratigraphic studies. Logan and Brown (1976) at Exmoulh Gulf describe a regional framework tor the coastal environment b\ delineating large-scale units termed geologic- physiographic provinces based on hinterland characteristics, rhercafier. at a smaller scale. ihc> identify various terrain and tidal flat units. Ihcsc units are described in detail and divided into a range ot smaller scale units on the basis of substrate, creek incisions, finc-scale bcdlorms and biota, fhe units include types such as low-intertidal zone, mid-intertidal zone, supratidal zone, tidal creeks, beach ridges, etc. In a siud\ along another pan of the Western Australian coast at Dampicr Archipelago. Semcniuk e/ al. (1982) provide an hcirarchial classification o( a coastal zone as a framework for further studies on biology and scdimenlology. At the largest scale four major marine settings were recognised: (I) Oceanic Zone. (2) Dampicr .Archipelago. (3) Nicko! Bay Comple.x, and (4) Maitland Delta Complex. Thereafter the paper concentrates on the Dampicr Archipelago and subdivides it into (gco)morphologic units such as submarine plains, islands, reefs and shoals, and (inter- related) channels, straits and cmbaymonls. These units arc further subdivided into smali-scalc "geomorphic units" on the basis ofgcomciry. substrate and tidal level, (c.g. inlcrlidal beaches, inlcrildal flats, intertidal rocky shore, etc.). Since the primatw objective of that paper was to describe the framework for biologic systems in the area, the next subdivision is termed a "habitat" ano units such as intertidal Hat are subdivided into a profusion of small-scale units useful tor biologic purposes. Jennings and Coventry (1973). Jennings (1975), and Thom ci al. ( 1 975) describe various geomorphic features in King Sound and Cambridge Gulf, respectively. Jennings and Covcntr> deal with the stratigraphic relationships and origin of small-scale spits and "barrier islands" along the eastern shore of King Sound. Jennings (1975) describes the stratigraphic relationship between Quaternary tidal flat deposits and red .sand dunes: Jennings also presents several generalised geomorphic profiles across King Sound tidal flats within which arc recognised three tidal-level zones and various tidal landforms such as sand shoals, cheniers. cliffs and lagoons. Thom et al. (1975) in a study of mangrove ecology in Cambridge Gulf similarly provide several generalised geomorphic profiles within which they identify three lidal-lcvel zones as well as beach ridges and creeks. In a series of papers on mangrove-lined tidal flats ot northwestern Western Australia. Semeniuk (1980. 1981a, 1981b. 1982. 1983. 1984) provides a subdivision and classification scheme specifically of tidal zone systems. Probably ilie most relevant paper to this study is Semeniuk ( 198 lb) wherein tidal zones arc subdivided. dcscribcd and mapped, and a classification of tidal Bat types presented based on substrate, stratigraphy, suites of geomorphic units and inferred Holocene history. Generally in all the work by Seineniuk {op cil.), the approach adopted was: (1) identification of geometric forms on the tidal zone (e.g. ridges vs flats v.s creeks). (2) recognition of slope (c.g. flats, slopes and cliffs), (3) identification of substrate types, and (4) identification of small-scale surface morphology (c.g. smooth surface such as salt flat v 5 hummocky burrow-mounded surface such as mangal flat). In this manner tidal flats were subdivided into salt flat, manga! flat, low tidal Bat, sand Bat. shoals, alluvial fans. Discussiou There arc several main conclusions that can be drawn from the literature on the central wpi. northwest and north coast of Western Australia. Firstly it is obvious that there has been a predominance of studies on deposuiona! areas such as tidal Bats and deltas, and few — if anv — on the other diverse geomorphic entities such as barrier islands, rocky shores, beach/dunc shores etc. Overall, the works on Shark Bay. Dampicr Archipelago and tidal Bats generally, serve to show that 57 Journal of the Royal Society of Western Australia, Vol. 68. Part 3, 1986. the heirarchial system of classification employed elsewhere in the world has been successfully applied in Western Australia though there is an inconsistency in terminology at the smaller scales, differing concepts of what constitutes the largest scale of reference in defining large-scale units, and some inconsistency in the use of criteria at all scales. For instance, the criteria on which large-scale units are recognised arc: (1) regional geology and physiography of the hinterland (Logan and Brown. 1976) ; (2) geomeli*}^ of coastal form (Semeniuk et al. 1982): (3) erosional w-tsus dcpositional system (Davies 1977) ; (4) regional processes (Davies 1977). The smaller scale units present yet another problem because they have been identified and subdivided yariouslv on numerous criteria that include geometry, slope, level relative to MSL and substrate. Few studies have attempted to come to terms with cither a nomenclature ora philosophy of approach that explicitly deals with the various scales of geomorphic units. It is also obvious that since the various authors have worked in diverse coastal systems, terminology has evolved for specific areas. This terminology is not applicable throughout the region. For instance, consider the example of tidal flats (Brown and Woods 1974, Hagan and Logan 1974. Jennings 1975. Thom et al. 1975. and Semeniuk 1981b). These authors have used a wide variety of criteria to subdivided tidal flats and hence develop independent systems of terminology. Brown and Woods (1974), Hagan and Logan (1974), and Logan and Brown (1976) utilise tide levels; Jennings (1975). and Thom et a.!. (1975) utilise tidal levels, substrate and slope, while Semeniuk (1981b) employs criteria of tidal level, slope, shape, substrate and small- scale morphology. A similar comparsion of terminology and criteria for subdivision for rocky shores (cf. Read 1974 and Semeniuk el al. 1982) also shows variability in approach and nomenclature. The same principle applies to other small-scale geomorphic units. In summary, it may be noted that authors tend to subdivide geomorphic entities in smaller units on w'hatcver criteria arc suitable or relevant to their particular study. These criteria of course are not consistent from discipline to discipline and consequently independent studies lend to result in a profusion of dissimilar terminology. There is therefore no single nomenclature system considered adequate for the whole region, but where established terminology is adequate or relevant to this paper, it is utilised latcr'on. Overall, however, it seems preferable to develop a consistent and new approach and terminology for the coastline of this study area. The proposed approach and terminology are discussed below. The Proposed Classification and Terminology: Use of Scale The purpose of this section of the paper is to rationalise the terminology and classification of tropical Western Australian coasts with particular reference to scale. This is approached in two ways: firstly, by reviewing the use of the term “geomorphic unit" and secondly, by proposing scalar terms for description/nomcnclaturc of various geomorphic features along the coast. The Term “Geomorphic Unit" One fundamental problem in many classification and terminology systems is the use of the term “geomorphic unit" or some" other equivalent term such as “facet" (cf. Bourne 1931. Brink et al. 1965). Most authors appear to use these terms at one scale only: thereafter, when referring to smaller or larger scale units, terms such as “elements" or “system", respectively, are introduced. When detailed studies proceed beyond the currently defined scalar frames of reference, terms are borrowed from related disciplines (such as sedimentology). To illustrate this point of scale-determined nomenclature, an example is drawn from work on the Swan Coastal Plain. Although outside the study area of this paper it serves to show how the terms “geomorphic unit'V“geomorphic element" arc utilised. The term “geomorphic unit" is used to refer to the Swan Coastal Plain itself and the term “geomorphic clement" is then used to refer to units within the Swan coastal Plain (McArthur and Bettenay ( 1 960) after Woolnough ( 1 920). If workers require to subdivide the “geomorphic elements" into finer scale catergories such as ridges versus swales, on current practices there arc presently no terms for the nomenclature for the smaller scale categories. This pattern of introducing new category terms for landform entities at each scale of reference is discussed in Brink et al. (1965), Perrin and Mitchell (1969) and Mabbutt (1968), and is a result of geomorphologists attempting to develop both a philosophy of approach and terminology concurrent with genetic classification. In practical terms, however, neither geomorphic units nor the aggregations (suites) of such units conform to any established size classes. Semeniuk et al. (1982) confronted similar problems in the Dampier Archipelago. Once the term geomorphic unit was allocated to features at a particular scale, then by principle of exclusion larger and smaller scale features could no longer be termed “geomorphic units". Semeniuk ei al. (1982) then referred to larger scale units as “morphologic units" and smaller scale units as “habitats". In reality all arc geomorphic units for their nominated scale. Semeniuk (1985) partly resolved this problem of geomorphic unit nomenclature by introducing scale terms to qualify the term “coastal features". Thus large-scale coastal ( geomorphic) features, medium-scale coastal (=geomorphic) features, and small-scale coastal features were described. If the use of the term “geomorphic unit" appears to be an obstacle to scalar classification and terminology then perhaps a discussion is required to determine if the term itself is a problem. The "geomorphic" component of the term refers to landform shape, and as such its meaning is reasonably explicit. A "unit" may be defind as the smallest entity recognised at a particular scale. Sand grains arc the units of a sand deposit at hand specimen scale, while embayments. inlets and rocky headlands are the units of a ria coast at the aerial survey scale. On this basis a geomorphic “unit" should be viewed as any recognisable or mappable landform entity within a nominated scale of reference. Ria coasts, deltas and rocky shores ma> be observable units at the regional scale while tidal flat subdivisions generally are not. However, the tidal flat subdivisions (units) become differeniiatcd at the medium- and small-scale of ob.scrvations. Thus, any landfonn within the various scales of reference may contain a set of observable units, and all of these should be termed “geomorphic units" as long as the scale of observation is nominated. 58 Journal of the Royal Society of Western Australia. Vol. 68. Fart 3. 1 986. It is proposed therefore that the term geomorphic unit be retained throughout descriptions of terrain/coastal zones but that the scale of reference he fixed and nominated. This allows a worker to describe features of a land surface to a level as fine or as large as is desired. This scalar approach is already utilised by oceanographers who refer to macro, nieso and micro- scale oceanographic features; by geologists who utilise macro, meso and micro-structural features (Turner and Weiss 1963); and by climatologists (Barry 1970, Barret 1974). Each of these disciplines, however, has its own concepts and boundaries of scale to which they refer macro-, meso-, and micro-. A reconnaissance of many standard geomorphology texbooks. however, will find scalar terminology or its equivalent generally missing form their index and contents (text). (Bird 1976, Bloom 1978. Embicion e! al. 1978, Davies 1980, Gardiner and Dackombe 1983. Gardner and Scoging 1983. Goudie 1981. King 1966, 1972, 1975, McCullagh 1978. Trewarlha ct al. 1968. and many others.) In contrast, where a scalar approach in terrain description is utilised by geomorphologisls, the hicrarchial classification (system, facet, element) is based on criteria of genetic relationships of landform units as well as scale {Linton 1951, Brink c/ i?/. 1965, Perrin and Mitchell \969): scale is not utilised in these studies as the sole framework. I'he Propo.sed Scale Terms The terminology proposed for the various scales of features evident along the tropical Western Australian coastline is as follows (Fig. 1 and Table 1); • Regional • Large • Medium • Small • Fine. Table 1 Summao labic of scale terms and their respective scales of reference Scale terms Frame of reference Regional (Megascale) scale 500km X 500km to 100km x 100km Large (Macroscale) scale 50km X 50km to 10km x 10km Medium (Mesoscale) scale 5kmx5kmtolkmxl km Small (Microscale) scale 500m X 500m to 10m x lOm Fine (Leptoscale) scale 5m X 5m to 1 m x 1 m Workers who prefer to use ancient Greek in the construction of terms may use Megascale, Macroscale. Mesoscale. Microscale and Leptoscale (see Liddell and Scott. 1925-1940 for definition of mega, macro, meso, micro, and Icpto) as synonymous terms. A description, with examples, of these scalar frames of reference is presented below-. Regional scale (or Megascale): morphology evident or mappable at the scale of a region, i.c. within frames of reference of 500km x 500km down to 100km x 100km. This scale would incorporate the term “land region'* by Linton (1951), Brink et al. (1965), and Perrin and Mitchell (1969). and would be termed "regional” by numerous other authors (e.g. Cooke and Warren 1973). The term "regional** as utilised here refers only to the particular size; other authors tend to use the term "regional” with genetic implication (e.g. Jennings and Mabbutl 1977. and Mabbutt 1968). Some examples of coastal types along the tropical Western .Australian coastline within this scale of reference arc; ria shores, delta lands, and bcach/dunc shores. Large scale (or Macroscale): morphology evident or mappable at frames of reference of 50km x 50km down to 10km X 10km. This scale would incorporate the term "land facet” by Linton (1951), Brink et al. (1965). and Perrin and Mitchell (1969). and perhaps would be termed "basin scale” by Cooke and Warren (1973). Examples within a ria coastal setting in northwestern Australia are (after Semcniuk 1985): riverine channels, narrow embayments, broad embayments. cliff/rocky shores, sandy shores, islands, and subtidal reaches or waterways. Medium scale (or .\fesoscale): morphology evident or mappable at frames of reference of 5km x 5km down to 1km X 1km. This scale would incorporate the term "site” bv Linton (1951), "land element” by Brink el al. (1965) and Perrin and Mitchell (1969). Examples within broad embayments of a ria coastal setting arc (after Semcniuk 1985)': spits. Cheniers, rocky headlands, tidal flats, tidal creeks and alluvial fans. Small scale (or Microscale): morphology evident or mappable at frames of reference of 5()0m x 500m down to lOm X 10m. This scale would still incorporate the terms "site” and "land element” by Linton (1951), Brink et al. (1965). and Perrin and Mitchell (1969). and would be termed "local scale” by Cooke and Warren (1973). Examples on tidal flats in northwestern .Australia arc: a smooth salt-cncrustcd mud surface (= salt flat); a smooth rippled sand surface (= sand flat); and a hummocky, burrow-mounded mud surface (=- mangal fiat). Fine scale (or Leptoscale): morphology evident or mappable at frames of reference of 5m x 5m down to 1 m X Im. This scale would incorporate the term "microrelier* by Hunt (1972). and “microform** by Tricari ( 1 972). Examples on tidal Bats in northwestern Australia include ripple marks, erosional rills and burrow mounds. For purposes of this paper there is no need to proceed beyond the fine scale. If frames of reference smaller than “fine scale** were to be utilised then the observations would be out of the realm of traditional geomorphology; thus fine-scale represents the lower scalar limit of the science of geomorphology in this paper. At the other extreme, there are of course frames of reference that extend beyond "regional scale"; however, in tropical Western Australia the next scale-unit above regional scale (i.e. 1 000 km x I 000 km) is subconiinental and would incorporate the entire study area within which units such as Pilbara coastline Canning Basin coastline and Kimberley coastline would be the primary components. Al the subconiinental scale geological features such as cratons. blocks and basins exert a major influence on coastal form, and therefore perhaps the nomenclature of larger scale .systems should follow geological subdivision based on tecionic/struclural/lilhoiogic criteria, a conclusion also reached by Davies ( 1 977). 59 46670-3 Journal of the Royal Society of Western Australia, Vol. 68, Part 3, 1 986. It should be noted that the nominated scales may be applicable only to the northwest and north tropical coast of Western Australia. Elsewhere coastal features may be of a different magnitude of size-variation, and a redefinition of absolute values of regional-, large-, medium- and small-scale may be necessar>'. The Proposed Classification: Use of Geomorphologic Terms The pu^ose of this section of the paper is to identify and describe various geomorphic units along the coast of tropical Western Australia within the five defined scales of reference. Landforms thus may be described in progressively decreasing scales, and a coastal type can be classified as a geomorphic unit at a particular nominated scale (e.g. a sand flat on a tidal zone is a small-scale geomorphic unit, a tidal flat can be a medium-scale geomorphic unit, while the deltaic complex to which they belong may be a regional scale geomorphic unit (Fig. 2). Criteria Numerous criteria can be used to identify geomorphic units (see literature reviews) and these criteria are applicable at all scales: • depositional versus erosional system (in a long- term Quaternary geological context) LARGE SCALE A. REGIONAL SCALE DELTA SUBTIDAL PRODELTA FLATS STRAND PLAIN SMALL SCALE C. MEDIUM SCALE SUPRATIDAL RIDGE SAN^-.,. ^ i r TIDAL ^ LJ CREEK TIDAL FLAT SAND FLAT TIDAL FLAT. MUD FLAT Figure 2. — The various geomorphic units in a deltaic setting observable and mappable at 4 scales of reference. 60 Journal of the Royal Society of Western Australia. Vol. 68. Part 3. 1986. • geometry’ of landform (plan geometry, slope, relief) • morphologic features of surface, at various scales • substrate types, which can influence the development of surface morphology at all scales • dominant (gcomorphic) processes at surface, which also influence development of various morphologic features • landsurface position, that is location within a coastal system (e.g. interface between hinterland and tidal flat). Many of these criteria already carry an implication of variability of landforms: for instance* the fact that a coastline is constructional (e.g. a della) implies there are a wide range of medium- and small-scale associated geomorphic features (such as sand spits, channels and flats) that are extremely different to those developed along an eroding shoreline (e.g. cliff and boulder> shores). Some of the above criteria also encompass the genetic classificalions/impticalions of other authors. For instance, a marine-inundated tluvially-dissected coastal terrain, which is termed a ria, may be a primary criterion for some authors (Johnson 1919, Shepard 1963), but it may have been used with genetic implication; the criterion ‘geometry of landform’ proposed here, however, is non-genetic, but it will still serve to distinguish these types of shorelines (rias) from other shore types. Geomorphic Units of The Tropical Western Australian Coast There is a limited range of geomorphic units that occurs within each of the scales of reference nominated above, and each scale of reference tends to have a very distinct suite of units, especially at the smaller scale. The geomorphic entities in north-west and north Australia that are evident within the five scales nominated above arc listed below and arc described in Tables 2-5, and maps are presented in Figs. 3-7. This list is by no means complete, especially at the smaller scales, and further work may refine, or add to the terminology. It should also be noted that some geomorphic units can make an appearance at a numher of different scales, because of the size variation of such units. Salt flats in high tidal zones exemplify this; they are evident at regional scale (King Sound), as well as large scale through to small scale, where they can be merely small patches 25m^ in size. Figure 3. — Map showing study area and location of detailed sites illustrated in Figs 4-7, 61 Journal of the Royal Society of Western Australia, Vol. 68, Part 3, 1 986. Many of the terms utilised herein have been obtained from the global and local literature and are cited accordingly. However, for some of the (progressively) smaller scale units new terminology has been developed in this paper. The use of some established terms sometimes arc used differently to some authors working in different environments (c.g. the term ‘beach ridge'). Nonetheless, the definitions of the terms as used in this paper are presented in Tables 2-5, Readers familiar with studies in sedimentology will realise that at many scales terminology in geomorphology and sedimentology is synonymous. Both disciplines essentially deal with surface morphology and consequently they describe the same features. Table 2 Regional-scale Geomorphic Units Unit Description Examples Archipelago Group of islands; grades into ria shore Dampicr Archipelago Barrier island complex Narrow, shore-parallel limestone barrier ridges which bar and protect inlets, lagoons and tidal embayments Port Hcdland coastline Beach/dune shore .Strip of shore parallel coastal dunes with shoreline beach, beach ridges and foredune Eighty Mile Beach Delta lands Cuspate to deltoid lowlands at mouths of main rivers Dc Grey River delta Gulf complex Large embayment or inlet penetrating deep into the mainland; grades into lidal embayment Exmouth Gulf Ria shore System of bays and inlets of riverine origin cut into a rocky hinterland; grades into archipelago systems Kimberley coastline Rocky shore ('oast cut into a rocky hinterland but without marked development of inlets Cape Range western shore Tidal cmbaymenl (lidal land) Extensive tidally-inundaled embaymcni or inlet grades into gulf system Roebuck Bay Regional scale geomorphic units Archipelago Barrier island complex Beach/dune shore Delta lands Gulf complex Ria shore Rocky shore Tidal cmbaymenl Some of these units arc iniergradational: ria shores and archipelagos: gulf complexes and tidal embayments: delta lands and barrier island complexes. Examples of these units arc illustrated in figs. 3-7. Description and occurrence of the units are presented in Table 2. Large scale geomorphic units Alluvial fan Barrier island Beach/dunc shore Broad embaymcni Cliff/rocky shore Headland Table 3 Geomorphic Units at the Large-scale Unit Description Selected examples Alluvial fan Fan to deltoid to elongate alluvial deposit King Sound west shore; Pilbara coast between Onslow and Dampier Barrier island Narrow limestone or sand ridges which may be mantled by dunes, beach ridges, soils and tidal deposits; surrounded by water at high tide Fiiuicanc Is.- Port Hcdland area; Port Weld; north-east of Onslow Beach/dune shore Shore-parallel coastal dunes with accompanying beach ridges, foredune and shoreline beach Eighty Mile Beach Broad embayment Broad inlet or embayment; with permanent water on ail tidal levels; margins arc iidally exposed Kimberley coastline; sec Fig. 7A, 7B Clitl/rocky shore Coast cut into rocky hinterland; may be composed of cliffs, or bouldery slopes, or benches. clifTs and pavements; may contain local pocket beaches Cape Range western shore; Kimberley coastline Headland Rocky coast promontory which may be composed of cliffs, bouldery .slopes, benches or pavements Cape Range north lip Island Supraiidal landforms surrounded by waterway or lidal lands Cape Preston: West Intercourse Is.. Dampicr Archipelago Narrow embayment Narrow inlet, with permanent water on all tide levels; margins arc lidally exposed Kimberley coastline; sec Fig. 7A, 7B Riverine channel Narrow channel system that is the seaward extension of riverine channels Fortescue River; Turner River Shoals Hummocky, undulating, expansive sheets and mounds of sand King Sound central embayment /one (Fig. 6A) Strand plain Lowland composed of linear beach ridges and dunes separated by intervening tidal lands Turner River delta; De Grey River della; Ashburton River delta Tidal flat (tidal land) Tidally-inundated lowland West shore King Sound, see Fig. 6.A; Dampier Creek. Broome Tidal creek Tidal-water drainage/channel system that typically incises lidal flats King Sound, see Fig. 6A Island Narrow embaymcni Riverine Channel Shoals Strand plain Tidal creek Tidal flat (and in many cases, types of tidal flat) Some examples are illustrated in Figs 3-9. Description and occurrence of the units are presented in Table 3. 62 Journal of the Royal Society of Western Australia, Vol. 68, Part 3, 1 986. Medium scale geomorphic units This group can be recognised on criteria listed above. Location relative to MSL also is useful to note. The list includes: Alluvial fan Alluvial plain Barrier Island Beach Beach ridge Chenier Dunes' Fluvial channel Foredune Hinterland/tidal flat margin Lagoon Levee Nearshore bar system Rock island Rock pavement Rocky shore Sand island Shoals Spit Tidal Creek Tidal HaC Some examples are illustrated in Figs 3-7 and Figs 9-10. Description and occurrence of the units are presented in Table 4. 1 able 4 Medium-Scale Geomorphic Units Unit Description Selected example Alluvial fan Fan to deltoid to elongate alluvial deposit Fig. lOD .Mluvial plain Ribbon to sheet alluvial deposit not illustrated Barrier island Narrow limestone or sand ridge which may be mantled by dunes, beach ridges, soils and tidal deposits: surrounded by water at high tide Fig. 4B Beach Intertidal slope of sand or gravel developed on a strip along the shore of dunes, beach ridges, spits, etc. Fig. lOE Fig. 1 2D Beach ridge Shoestring sand (or gravel) deposit developed to supraiidal level by storm activity; occurs to landward of beach slope not illustrated Chenier Detached shoestring or bar sand deposit built to high tidal or supratidal levels surrounded by muddy tidal-lands; may be tidal to supratidal Fig. I OB 'Types of dunes, such as transverse, parabolic, linear and barchan can also be differentiated. Hn many instances, types of tidal flats such as salt flats, mangal flats and low tidal flats arc recognised, although the small distinguishing characteristics that comprise the photolone evident on an aerial photograph are not evident at this scale. Table 4 — continued Medium-, Stale CJcomorphic Unils Dunes Shoestring to lensoid to mound-like accumulations of sand of some relief developed along the coast by onshore aeolian activity; may be subdivided on external geometry and relation to wind direction (McKee. 1979) into linear, parabolic, /ramvmcand barchan types; dunes may be mobile or immobile, and bare or vegetated (.“Mso see foredune). Fig. 8C Fluvial channel Channel system of rivers which meet the coast Fig. lOD Foredune Shoestring deposit of sand developed by aeolian process usually as a low ridge immediately landward of the beach not illustrated Hinterland/ tidal flat margin Complex system of interlace between hinterland and tidal flats; ma\ be narrow or broad; diffuse to sharp Fig. 7C Lagoons Impounded depression or channel not illustrated Levee Narrow channel-paralled mound or rise developed on bank of channels not illustrated Low tidal to near-shore bar system System of low-relief bars and intervening troughs developed on low tidal to shallow subtidal zones not illustrated Rock island Supratidal island of limestone or sandstone or Precambrian basement surrounded by waterways or tidal-land Fig. lOB Rock pavement Extensive low-lying subhonzonial to gently-inclined pavement of rock (either limestone or sandstone or Precambrian basement) Fig. 10 E Rocky shore Shoreline composed of clitfs, or sleep slopes or bouldcry deposits; locally- developed pocket beaches Fig. 5B Shoals Hummocky to undulating sheets and mounds of sand not illustrated Sand island Supratidal hummock of sand surrounded by tidal lands Fig. lOA Spit Shoe.siring or bar sand deposit emanating from headland of rock or dune field; may be tidal to supratidal Fig. 9B Tidal creek Meandering to bifurcating to ramifying drainage systems cut into tidal flats; may drain out on a low tide Fig. lOF Tidal flat (and. in many cases, types of tidal flats; see Tables) Gently-inclined tidally-inundated lowlands Fig. IOC 63 Journal of the Royal Society of Western Australia, Vol. 68, Pari 3, 1 986. Table 5 Geomorphic Units at the Small Scale Medium-scale geomorphic setting Small-scale units Description Occurrence with respect to tidal level Alluvial fan channel drainage/distributary incision lobes progradational/accretionary lobate promontory at margins of fan depending on region, all units may be located anywhere between levels LWN to supratidal flat relatively flat surface of alluvial fan Alluvial plain channel drainage/distributary incision supratidal flat relatively flat surface of alluvial plain Bar system bars low relief sand wave low tidal to subtidal troughs intervening swale between bars Beach beach slope intertidal slope of beach intertidal: MLWS-MHWS backshore(= berm) impermanent nearly horizontal or land sloping bench on backshore of a beach storm water levels Beach ridge beach ridge crest highest line or surface of a beach ridge storm water-supratidal level beach ridge slope flank of a beach ridge high intertidal to suptratidal beach ridge swale trough between any 2 successive beach ridges hummock irregular mound on surface Chenier Chenier crest highest line or surface of a chenier high intertidal-supratidal chenier slope flank of a chenier chenier lobe accretionary lobate promontory at inner margin of chenier Dune dune crest highest line on surface of dune all supratidal dune slope flank of dune dune swale trough between any 2 successive dunes dune hummock low relief sand mound Foredune foredune crest highest line of surface of foredune all supratidal foredune slope flank of foredune foredune hummock low relief sand mound Fluvial channel channel water-filled or dry, relatively narrow erosional incision all supratidal bars/shoals moundlike sediment accumulations in mid-channel areas banks steep margin of channel Hinlerland/tidal flat margin gravel apron muddy sand to sand apron muddy sand to sand sheet narrow ribbon of sedimentary material bordering a supratidal area of bedrock, or limestone, sand plain; slope generally steeper than adjoining tidal flat but less so than hinterland generally high tidal-supratidal; in some cases mid-tidal to supratidal channels/gutters erosional incisions Levees (fluvial) crest highest line or surface of levee all supratidal slope inclined surfaces of levees gutters erosional channels cut into levees Rock island cliff vertical/steep rocky surface high intertidal to supratidal gravel/sand apron ribbon deposit of gravel/sand flanking island channels/gutters erosion incisions subaerial surface the varied subaerial surface of an island supratidal 64 Journal of the Royal Society of Western Australia, Vol. 68. Part 3. 1986. Table 5 — continued Medium-scale geomorphic setting Small-scale units Description Occurrence with respect to tidal level Rocky shore cliff shore vertical/steep sheer surface these units occur at various levels from supratidal, intertidal to subtidal fissured rocky shore vertical/sleep to inclined, guttered to cracked surface gutter erosional incision pavement flat to gently inclined surface bench narrow terrace gravelly shore gravel accumulation in sheet, ribbon or lens form bouldery shore boulder accumulation in sheet, ribbon or lens form pocket beach sand accumulation in lens or sheet form reef protruding knoll of rock Rock pavements limestone pavement flat to moderately inclined pavement of limestone low tidal to supratidal rock pavement flat to moderately inclined pavement of rock other than limestone, e.g. Precambrian rock cliff small cliffs usually 2m cut into the pavements pool depressions 1 m to several metres in size bench narrow terrace Sand island crest/top/plain highest surface of island supratidal slope flanks of island high tidal to supratidal sand flat apron ribbon of gently inclincd/flat sand deposit circumferential to island sand cliff small cliff usually 2m cut into sand at margin of island creek/gutter erosional incisions cut into islands Spit spit crest highest line or surface of a spit high intertidal-supratidal spit swale trough between 2 successive spits spit slope flank of a spit spit lobe accretionary lobate promontory Tidal creek channel relatively narrow erosional incision intertidal to subtidal bank steep-walled margin of creek intertidal levees linear, low mound-like sediment deposit bordering the margin of creeks shoal mid-channel mound-like sediment deposits intertidal to subtidal mouth fan fan-shaped accumulation of sediment at mouth of creek intertidal (to subtidal) point bar lensoid sediment accumufation on convex meander of creek Tidal flat low tidal sand to muddy sand flat flat surface underlain by sand or muddy sand low tidal low-mid tidal mud flat flat, smooth surface underlain by mud low-mid tidal gravel flat flat surface underlain by gravel low tidal, varying to high tidal salt flat flat smooth salt-encrusted surface high tidal mangal flat flat to gently inclined burrow-mounded surface vegetated by mangroves, underlain by mud, sand or muddy sand mid to high tidal shoal hummocky mound of sand low tidal slope gently inclined slope underlain by mud mid-low tidal cliff vertical/steep surface usually 2m high usually at LWN and HWN level shell pavement flat surface underlain by shell low tidal, varying to high tidal 65 CO 5 Q llJ ii lO <2 si- I ^ —I QD I < ^ Z3 < ^ cr o Q. CC 30- ZQ UJZ s- ^5 LiJ< 22 esp^ CO UJX <5! o< — Q ?CD 03 _JCO CD CO ‘='2 do CM < < ''v 2 LU CC < _l < UJ UJ cr 0 _i _l CO cr UJ 2 0 3< n — ' ^ or ISLAND < Li. _l < FLAT Cl < q: Q UJ > _j p Q CC 0 1- Q 3 < CD — < 0 2 2 _J 0 3 h- CD _i — < CO U. X CO < h- Figure 4. — Geomorphic units evident along a barrier island coast near Onslow. A. At regional scale, B. At large scale. C. At medium scale. 66 Journal of the Royal Society of Western Australia, Vol. 68, Part 3, 1 986. A. INSErE „Jn§et^c IN^SET G ^ NSET E INSET ROCKY H'LAND ROCK ISLAND ^ H'LAND/TIDAL FLAT ^ I MARGIN^ COLLUVIAL GRAVEL APRON □ SALT FLAT^(A HIGH TIDAL MUDDY SAND FLAT ) □ LOW TIDAL LIMESTONE PAVEMENT ROCKY SHORE GRAVEL SHORE (MANGROVE VEGETATED MID TIDAL SAND FLAT ROCKY HINTERLAND COLLUVIUM/ALLUVIUM APRONS AND RIBBONS DUNES BEACH ROCKY SHORE MANGALFLAT (A Ml D- HIGH ^ TIDAL SJRROWep MUD /SAND flat with MANGROVES) MID-LOW TIDAL FLAT TIDAL CREEK SUBTIDAL AREA □ ROCKY H’LAND ROCKY SHORE BEACH, BEACH RIDGE AND SPIT DUNES I (^\ SAND ISLAND I rN I ROCKY ISLAND I wl /ROCKY REEF TIDAL CREEK ALLUVIAL FAN H'LAND TIDAL FLAT MARGIN ^ COLLUVIAL AND BOULDER APRON SALT FLAT ( A HIGH TIDAL SAND FLAT ) MANGAL FLAT (A MID-HIGH TIDAL BURROW- MOUNDED MUD AND MUDDY SAND FLAT WITH MANGROVES ) MID-LOW TIDAL FLAT SUBTIDAL AREAS 1 FLOOD PLAIN / r“ I SAND PLAIN HINTERLAND c::) SAND ISLAND SALT FLAT TIDAL CREEK IT A ROCKY HINTERLAND ROCKY REEF SPIT MANGAL FLAT {BURROW-MOUNDED MUDDY SAND FLAT WITH MANGROVES) MANGAL FLAT (HUMMOCKY SAND FLAT With MANGROVES) LIMESTONE PAVEMENT ROCK GRAVEL PAVEMENT SHELL PAVEMENT SAND SHOAL RIPPLED SAND FLAT HUMMOCKY MUDDY SAND FLAT TIDAL CREEK CHANNEL TIDAL CREEK LEVEE SAND DUNES ALLUVIAL FAN BEACH □ SALT FLAT ( A HIGH TIDAL SAND FLAT ) MANGAL FLAT (A MID-HIGH TIDAL HUMMOCKY SAND FLAT WITH MANGROVES ) MID-LOW TIDAL FLATS Figure 5. — Gcomorphic units evident along an archipelago-ria coast. Dampier Archipelago. A. Regional scale. B and G, Medium scale. C. D. E and F. Small scale. 67 LlJ Oo _) z ^cc< X Xwtn COH >X I^cou. 200 q:<“ _|0 CJUJ -JZ zq: << o ;;V Figure 6. — Geomorphic units evident along a gulf. ICing Sound. A. Regional scale. A broad tidal cmbaymonl is outlined as inset B. B. Large scale. C and D. Medium scale. 68 Figure 7. — Geomorphic units evident along a ria coast. Port Warrender, Kimberley area. A. Regional scale. B- Large scale. C. Medium scale. D. Small scale. Small scale geomorphic units This list is quite large because many of the medium scale geomorphic units can be satisfactorily subdivided on slope, geometry, small-scale and fine-scale morphology of the substrate surface. Some examples are illustrated in Figs 1 1-13 and some are listed below, but a more comprehensive listing is provided in Table 5 along with definitions. Tidal flats as medium-scale geomorphic units may be subdivided into small-scale geomorphic units on criteria of slope, substrate type and fine-scale surface features (Figs 5D & 14). Some examples using tidal flat surfaces are: Gravel flat Mangal flat (•burrow'-mounded mud flat that is mangrove vegetated) Inclined mud slope Salt flat (=smooth, salt-encrusted mud flat) Sand flat Sand shoals Shell pavement Small cliff Smooth mud flat Tidal creeks lend, to be internally heterogeneous and may be subdivided into: Creek bank Creek channel Creek levee Creek mouth fan Creek point bar Creek shoal Dunes, foredunes, and beach ridges may be subdivided into: Crest Hummock Slope (or flank) Swale Journal oflhc Royal Society of Western Australia. Vol. 68. Part 3. 1986. Fine scale geomorphic units This list also is large because a variety of physical, chemical and biological processes inlcracl with small scale geomorphic surfaces to develop a profusion of products. Some examples arc illustrated in Figs. 11-13 and some are listed as follows: Burrow mounds (on sand, or mud); see Fig. 1 1C Burrow scours (on muddy sand); see Fig. 1 1 F Desiccation cracks (on mud); see Fig. I lA, 1 IB Erosional rills (on sand, or limestone) Honeycomb surface (on limestone) Imbriccated gravel pavement Platcy gravel pavement Megarippics (on sand) Micropinnacles (on limestone) Ripples (on sand); see Fig. 1 1 D Scour marks (on sand or mud); see Fig. 1 1 A 1 IB Small cliff (cut into mud flats) Much of the variability at this scale can be related to differences in substrate and types of processes. For instance rocky shores cut into igneous rock will develop a suite of Fnc-scale features that are different from those B 70 Journal of the Royal Society of Western Australia. Vol. 68. Part 3. 1 986. Figure 8- — Some e.Naniples of coONlline cv idem al regional scale, A. Tidal erahaymem (Roebuck PUiins^'Rocbuck Hu\). B. Beach/dunc **luiro (lU'ai Onslow), f Barrier island cojiiplex near Port V\eld showing (1) bmestonc barrier island which is bordered to seaw-ard by (2) mangal flat and (3) low tidal limestone pavement and sand Hat; the barrier island protects a tidal embayment within which arc evident salt flats and (5) (mangrove-lined) tidal creeks. formed where rocky shores are developed on shale or quartzite (Davies 1980). As a result a separate list of fine-scale rocky shore morphologic features could be compiled virtually for cver\' unique gcological- iithological system that is sol in Ihc various oceanographic, chemical and biological sellings. Fine- scale morphologic features on sedimentary surfaces present yet another problem in variability. While there may be a greater tendency for sedimentary surfaces to portray a recurring pattern of limited number of bedforms (c.g. ripples arc ripples regardless of whether they are developed on fine calcareous sand, medium siliceous sand or coarse lithoclastic sand along the Pilbara, Canning Basin or the Kimberley coastlines), there is the factor of dynamics and temporal variation. Yesterday’s plane sand flat may. through spring tide action or'slorm acti\ ity. become today's rippled shoal. Compiling a list of fine-scale features would not be useful and relevant al this stage. The list would be very incomplete, and it probably would be best left to individual workers to identify the various fine scale features of a shoreline at their particular study sites. Use of tidal terms It should be noted that tidal level is not considered a primary criterion in distiguishing small-scale gcomorphic units. Nonetheless it may be used to locate particular portions of a tidal gcomorphic unit relative to MSL. Consider smooth mud fiats for example (Fig. 14). Smooth mud flats occur either above high water spring tide as firm, sall-cncruslcd, desiccated surfaces (- a salt flat), or al about low water neap tide; the latter is burrow-pocked and thixotropic. It seems preferable to distinguish between the two by referring to their tidal level or to some other conspicuous feature (such as salt encrustations, or burrows) rather than setting out a string of adjectival descriptors as a prefix viz. smooth, desiccated, salt-encrusted mud fiat. Thus two mud flat types may be distinguished by their relationship to tidal level, c.g. high tidal mud flats (or salt fiat), and low tidal mud flats. It is suggested therefore that in instances where a medium- or large-scale tidal gcomorphic unit can be subdivided on the basis of small-scale and fine-scale features but where the adjectival prefixes become loo cumbersome, the small-scale subdivisions should be indenlified by tidal level. Even if a small-scale gcomorphic unit is distinct in terms of its nomenclature {c.g. grave! fiat) and would not be confused with similar adjoining units, then a tidal level description could still be used at least to locale the unit relative to MSL. The tidal level description however is not a morphologic feature nor a gcomorphic subdivision, but merely identifies where a particular gcomorphic unit is occurring. In some cases distinctive gcomorphic units with distinctive small- and medium-scale features occur in a wide variety of geographic localities and recur in a specific pattern relative to MSL. Sall-encrustcd, smooth mud flats occurring above levels of mean high water spring tide and burrow-mounded, mangrove-vegetated mud flats occurring between mean sealeycl and mean high water spring tide exemplify this. Since these are inherently distinct units, they may be distinguished by their conspicuous features and termed “salt flat" and "mangal fiat", respectively. Flowever, some workers may prefer to use high tidal, smooth mud fiat and mid tidal, burrow-mounded mud flat, respectively, for these units. 71 Journal ofthc Royal Society of Western Australia, Vol. 68, Part 3, 1 986. Si' '*■ Figure y. — Some examples oCan archipelago-na shore. A B. ^Archipelago^^*^"^^* showing broad eniba>meiUs. wnh marginal tidal Hats, and siraiis/channels; width oJ view in background is 10km. Dampier Uml’iion O ^ubt.dal zone, (2) low-m.d t.dal flat, (3) mangal Oat, (4) sal, flat, (5) spits. 72 Journal of the Royal Society of Western Australia. Vol. 68. Part 3, 1 986. Figure 10. — Examples of geomorphic iiniisevidem ai mcdiiini scale in a vancty of coastal settings. A. (1 ) sand islands and |2) tidal creeks surrounded h> sail flat near Onslow: width of view 2km. B. Rock islands (arrowed), protruding through salt flat, Mitchell River estuar>', Kimberley. C. Channelled low-mid tidal mud slope succeeded to landward by mangal flat, chenier (arrov -j and salt flat. King Sound. Width of view is approximately 1km. D. Coast showing (I) barrier island, (2) beach ribbon, (3) alluvial fan. (4) mangal flat, and (5) riverine channel; Fortescue River. Width of view is approximately 1km. E. Coast showing (1) low tidal .sand flats. (2) h.-ncsione pavement, (3) mangal flat. (4) high tidal sand flat. (5) beach, and (6) supraiidal barrier island; near Onslow. Width of view is approximately I km. F. Tidal creek showing steep creek banks and niangrove-vegeiaicd mid-crcek shoals. King Sound. Width of view is 3km. Geomorphic Units and Habitats The term "habilal" refers to space in which abiotic factors determine as suitable for colonisation by biota, and a geomorphic approach in describing habitats merely identifies many of the major attributes of an environment that are critical to maintaining or eliminating elements of the biota. For example, landforni and substrate may control the variability, stability or dynamism of a shoreline; the type of substrate may have its effect on biota through mobility, permeability, transmissivity, nutrient/food retention, oxygenation, etc. A system of geomorphic units therefore forms a logical framework for the delineation/ identification of habitats. 73 Journal of ihc Royal Society of Western Australia, Vol. 68. Part 3. 1986. Figure I ! . Examples ot'fmc-scule gcomorphic units. A. Scoured- smooth mud Hal surface \Mih desiccation polygons on a sail flat. King Sound, B. Variety of gcomorphic features in a small tidal creek cut into a salt flat. King Sound. Scale is 30cm long. C. Hummocky, burrow-mounded surface on a mangal flat, Dampier Archipelago. Width of view is Im. D. Smooth, burrow-pocked mud flat separated by small clifT from a rippled sand ribbon. King .Sound. E. Small clifl. ZOcm high, and breccia deposit, cut into salt flat. Dampier Archipelago. Hammer for scale. F. Hummocky, low-udal. muddy sand flat. Dampier Archipelago. W'ldlh of foreground is 1 Om. Several authors have already utilised a gcomorphic framew'ork as a basis for ideniification of habitats (Thom 1967. Phicgcr 1977, Semeniuk el ai 1982). Also, in many biological treatises, the notion of "habitat" is rooted deeply in. or overlaps with, gcomorphic concepts (eg. Eltringham 1971. Yongc 1966, Odum 1971) and essentially these works implicitly identify the obvious (geo)morphology of an area and term 'such features habitats. This is not surprising considering that benthic j organisms interact intimately with the shape, type and dynamics of the substrate. In this paper at each scale of reference listed above. Ihe term gcomorphic unit in practical terms is interchangeable with the term "habitat" when a particular landform type is identified. For instance. 74 Journal ofihc Royal Society of Western Australia. Vol. 68. Part 3. 1986. Figure 12. Examples of juxtaposition of small-scale geomorphic units evident in seriical aerial photographs m Dampter Arcliipelago Fine-scale variation between units is also evident in some photographs. A. (1) Hummocky, low-lida). muddy sand tlal. (2) sand shoal locally vegetated by mangroves, and (3) tidal creek. Width of view is 100m. B. Tidal creek with components of ( 1 ) channel. (2) shoals. (3) levees: the creek traverses a hummocky, low-tidal, muddy sand Oat. (4) and locally a rocky reef. (5) preludes. Width of view is 100m. C. Low-tidal zone within which is evident ( 1 ) smooth muddy sand flat. (2) a tidal creek and (3) a smooth sand shoal. Width of view is lOOm. D Rocky hinterland ( I ). bordered by a beach ribbon of sand (2), and an intlincd rocky shore (^) within which arc osident \'anous tine- scale variations, the low tidal Hats are noted as (4) Width of view is lOOm rocky shores may be mapped as a regional- lo medium- scalc geomorphic unit and at these scales, rocky shores also may be viewed as a particular habitat for a range of organisms. Thus habitats may be viewed in a decreasing scale similar to geomorphic units, until at the smallest scale the biologist deals with “microhabilat” which is perhaps equivalent lo, but may be smaller than the fine- scale geomorphic unit. To illustrate this principle consider again the rocky shores {Fig. 13). .\l the small scale this habitat type may comprise cliff shores, bouldcry shores, sloping shores, pocket beaches, in which various tidal levels can be recognised as subdivisions of the rocky shore. M still finer scales exposed shear surfaces, notches, gravel accumulations, fissures and benches provide even smaller scales of reference for habitats. The only complication in relating habitats lo geomorphic units is that at some stage similar geomorphic units may be exposed to differing physico- chemical conditions and so would be different habitats. Rocky shores inundated by hypcrsaline water are a different habitat to those inundated by oceanic or brackish water. However, other factors being equal, purely on surface forms and features, geomorphic units may be equated with habitat units as long as the scale of reference is nominated. Discussion The results of this review and the proposed classification arc directly applicable lo the coast of tropical Western Australia since the philosophy was mainly developed on a data base from that region. However, the same approach, if not the detailed terminology, can be applied to other marine environments and other tracts of coast along Western Australia. For instance the deeper water sublidal shelf environments of tropical northwestern .Australia, and the coastal region of southwestern .Australia where the present Quindalup and Spearwood dune systems form continuous shoreline belts may he similarly classified utilising the approach presented here. .tckfuiwU’Ji^ements — the manuscript was cnltcally read by I). K. Cilasst'ord. IJ. J. Searlc and F. J. Woods, who provided useful discussion and commentary. Their help is gratefully acknowledged. 75 Journal of the Royal Society of Western Australia. Vol. 68, Part 3, 1986. Figure 13. Variahiluy at the small and fine scale along 2 shoreline types. A limestone bamer-i.sland shores bomeen Pori Hedland and Onslow. B and C are rocky shores along the Dampier Archipelago. D. E and F are A. Rocky shore showing cliff headlands alternating wnh houldcry shores. Field of view in foreground is 5m wide. B. Rocky shore composed of sheer cliffs, fissured cliffs and boulders. Field of view is .^m wide. C. Rocky shore composed offl.ssurcd slopes inclined towards right, alternating with steep/vcrtical Ussured cliffs. Person (arrowed) for scale. 2m Sgh"*’’ microscale hummocks and local areas oi microscale pinnacles m centre of field. Trees tor E. Limestone shore at mid-tidal zone showing pinnacles developed on top of an elongate reef. Field of view is approximately 1 Om wide. F. Limestone shore at high-tidal level showing 5m high cliff with pinnacles and boulders developed on surface. Person for scale. 76 Journal of ihe Royal Society of Western Australia, Vol. 68. Part 3, 1986. References Allen. J. R. L. ( 1 970).— Sedimontfi of the modern Niger Delia; A summarv- and review; in Morgan . 1 . P. {ed.K Deltaic scdimcnlaium: modern and ancient, hcon. Falcon!. Vfincralogisi.s SFtX' Puh! . 15: 138-151. Barret. E. C. (1974). — Ctimaioh^v from saicllitci. Methuen. London, 418p. Barry. R. G. (1970).~A framework for climatological research with particular reference to .scale concepts. I'ranmctions. Insli(ute oj British (ieographers. 49: 61-70. Bales. R. L.. and Jackson. J. A. J., 1980.— CrA/s.van- of geology (2nd Ed). Am. Geol. Institute. Falls duirch. Virginia; 749p. Bud, E. C. F. (1976a ). — Coasts (2nd Ed.), Australian National University Press. Canberra. Bird. E. C. F. (1976b). — The nature and source of be^ch materials on the Australian coast: in Davies. J. L. and Williams. M. A. J. (eds). Landform Kvolulion in Aiisiralia. Australian National University Press. Canberra: 144-157. Bloom. A. L. (1965). — The explanatory description of coasts. Zeil. fur Geomorph., 9: 422-436. Bloom. A I. (1978), — Geoniorplh'togy. J Systematic Anaivsis of Late Ccnitzoie Ltindfonn's. Prentice Hall. New Jersey: 5 1 Op. Boothroyd. J ('. (1978).— Mcsolidal inlets and estuaries, in Davis K. A. Jnr (ed ). ('oaiuil Sedimcntarv IGniroimicnis. Springcr-Verlag. New York: 287-360 Bourne. R. ( 1931). — Regional survey and its relation to .stocktaking of the agricultural resources of the British Empire. Oxford Fore.stry Memoirs. No. 13. Brink. ,A. B, A,, Mabbuit .1. A, Webster. R.. and Beckett, P. H. E. (1965). — Report of the working group on land dassitlcation and data storage )//: \E Kept No 940 Brown. R. G.. and Woods, P. J. (1974). — Sedimentation and tidal-flat devdopmeni. Ntieniah Embayment. Shark Bay Western Australia; in Logan. B. W., a al.. Evolution and l>iagenesis of Quaternary Carbonate Sequences. Shark Bay. WcMern Australia. Am. A.ssoc. Felrolcuin tteol. Mem . 22: 316-34(1. Chapman. V. J. (1976). — Mangrove I'egetation. J. Cramer; 447p. 77 Journal of the Royal Society of Western Australia. Vol. 68. Part 3. 1986. Coleman. J. M., Gagliano. S. M.. and Smith. W. G. (1970). — Scdimcniaiion in a Malaysian high tide tropical delta; iti Morgan. J. P. (ed.). Deltaic Sedimentation; Modern and Ancient, Soc. Earn. Palconi. Mtneralofiisi Spec. Puhl., 15; 185-197. Coleman. J. M.. and Wright. L. D. (1975). — Modern river deltas: Variability of process and sand bodies: in Broussard. M. L. (ed.). Dellas: Models Jnr Kxphranon. Houston Geol. Soc.; 99-149. Cooke. R. V.. and Warren. ,A. ( 1 973). — (k’omorpholoiiy in deserts. Balsford. London; 394p. Cooper. W. S. (1958k — Coastal sand dunes of Oregon and Washington. Oeol. Soe im. Mem . 12: l(>9p. Colton. C. A. (1942). — Shorelines of tranverse deformation. ./. Gcomorph.. 5:45-58. Colton, C. -A- (1952). — Criteria for the classification of coasts. I7th Ini. Geol.Conp Abs of Papers. 15. Davies, ,). L. (1964). — A Morphogenetic Approach to World Shorelines. 7.eil. fur Gcomorph., 8: 127-142. Davies, J. L. (1977). — The Coast; in Jeans, D. N. (cd.). Australia, a ^en^raphy. Sydney I'nivershy Press: 134-151. Davies, J. L. ( 1980). — Geographical vunulion in coastal development (2nd ed.). Gcomorphologv Text 4, l.ongman. New Y'ork; 212p. Davis. R- .A. ( 1978).— Com^ti/ Sedimentarv Environments. Springer-Verlag, New York; 420p. Davis. R. A. Jnr (1978). — Beach and Nearshore Zone; in Davis. R. A. Jnr (ed.), Coastal SedimentarY Environment.':. Spnnger-Vcrlag, New York; 237-285. Dav. J. H. (ed.l (1981 ) — Estuarine EcolofP'. A. A. Balkema. Rotterdam; 41 Ip. Dyer. K. R.. (1973). — Esiuanes: a physical introduction. W'iley and Sons, London; I4()p. Eiinngham. S. K. ( 1 97 1 ). — Li fe in mud and .sand. Linibooks. London: 2 1 8p. Emblcion, C'.. Brunsden, D., and Jones. D. K. G. ( 1 978). — Geomorphology. Present problems and future prospects. Gxford Univ, Press. Oxford: 281p. Evans. G,. Murray. J, W.. Biggs. H. E. .1., Bail., R.. and Bush. F. R. (1973). — The oceanography, ecology. sedimentology and geomorphology of parts of the Trucial Coast Burner Island complex. Persian Gulf: in Purser, B H (ed ). fhc Persian Gulf Springer \cr\ag, Berlin: 2.53-277. Fairbridgc. R. W. (1950). — Recent and Pleistocene Coral Reefs of Australia. J. Geoi, 58: .330-401 Fisk. H. N. ( 196! ). — Bar-llngcr sands of the Mississippi Della; in Geometry o! Sand Bodies. Am. Assoc. Petroleum Geol., Tulsa, f)k!ahoma: 29-52. Fisk. H. N.. McFarlan. E. Jnr, Kolb, C. R., and Wilbert. L. J, Jnr (1954). — Sedimcniaiy framework of the modern Mississippi Delta. J. Sed. Petrology. 24 76-99. Frazier, D. E. (1967). — Recent deltaic deposits of the Mississippi River; their development and chronologv. (/n// Cot/.sv Ivvo<' Geoi Soc. Trans.. 17:287-515. Galloway. R W.. (1982). — Distribution and Physiographic patterns of Australian Mangroves m B. F. Clough (Editor). Mangrove Eco.sysicms in Australia: Structure, function and management. AIMS and ANU Press, Canberra: 31-54. Gardiner. V.. and Dackombe. R. (1983). — Geomnrphological field manual. Allen and IJnvvm. London: 254p (iardner. R., and Scoging. H. (1983). — Megageomorphology. (’larendon Press. Oxford 240p. Geological .Survey of Western Australia (1980). — Western .Australia 1:50 000 Urban Geology Series. Dampier-Eaglehawk Island-Rosemary Sheet 2256-4 part of sheets 2 1 56- 1 and 2257-.5. First edition. Geological Survey of W'estern Australia (I982a). — Port Hedland-Bedoul Island. Western Australia 1:250 000 Geological Series with Explanatory Notes Geological Survey of Western Australia (1982b). — Onslow. Western Australia 1 :2.5b 000 Geological Series with Explanatory Notes. Geological Survey of Western Australia (1982c). — Madora, Western Australia 1 :250 000 Geological Scries w ith Explanatory Notes. Ginsberg, R. N, (ed.) (1975). — Tidal depcnifs. Springer-Verlag. Berlin: 428p. Goldsmith, V. (1978). — Coastal dunes: in Davis. R. A. (ed.). Coastal Sedimentarv Environments. Springer-Verlag. New ^’ork: 171-235. (ioldsmith, V.. Henningar, H. L, and Gutman. A. L. (1977). — The ‘ VAMP” coastal dune dassil'icaiion; in Goldsmith, V. (ed.), Coastal proces.ses and re\ulling forms of sediment accumulation, Currituck Spit. Tirginia/Sorth Carolina. SRAMSOE No 14 5, Virginia Institute ol' Marine Science. Gloucester Point. Virginia: 26-1 to 26-20. Goudio, A. (cd ) ( 1 98 1 ), — Geomnrphological technupie^ Mien and L^nwin. London: 395p. Ciould. H- R (1970). — The Mississippi delta complex; in Morgan. J P. (cd,). Deltaic sedinieniaiiun; modern and anciciil. Soc. Econ. Paleont. and Mineralogists, Spec. Pubi, 15: 3-30. Hagan. G. M., and Logan, B. W. (1974). — History of Hutchison Embaymcnt Tidal flat. Shark Bay. Western Australia: in Lo^n. B. W., el al.. Evolution and Diagcncsis of Quaternary Carbonate Sequences, Shark Bay. Western Australia. Am. l.rsoc. Petroleum Geologists Mem. 22: 283-315. Hayes. M. O. (1975). — Morphology of sand accumulation in estuaries; in Cronin. L. E. (ed.). Estuarine Research I'oL 2, Geology and Engineering- Academic Press, New York: 3-22. Hayes. M. O., and Kana, T. W. (1976). — Terrigenous clastic dcpositional environments. Tech. Kept No. Il-CRD, Coastal Research Division. Dept of Geology. Univ. .South Carolina; 302p. Hesp. P. .A- and Craig. G. F. (1983). — Sand dunes, storm ridges and Cheniers; in Craig. G. F. (ed.). Pilbara Coastal Flora. Western Australian Dept Agriculture; 7-14. Holmes. A (1965 ). — Principles of physical geology (2nd ed.). Nelson. London; I288p. Hunt, C B. (1972 ). — Genhgy of soils, their evolution, classification and uses. Freeman and Co.. San Francisco; 344p. Inman, D L.. and Nordstrom, C. E. (1971). — On the tectonic and morphologic classification of coasts. X Geoi. 79: 1-21. Jennings. J. N.. 1975: Desert dunes and estuarine fill in the Fil/roy estuary. Norih-wesiern Australia, Catena. 2: 215-262. Jennings. J. N.. and Bird. F. C. F (1967) — Regional gcomorphological characteristics of some Australian estuaries; in Lauff. G. H. (ed.). Estuaries. Am. Assoc. Adv. Sci.; 12I-I28. Jennings. J. N. and Coventry. R. J (1973). — Structure and texture of a gravelly barrier island in the Fiizroy estuary. Western .Australia and the role of mangroves in shore dynamics. Marine Geoi. 15: 145-168. Jennings. J N . and Mabbuil. J. A. (1977). — Physiographic outlines and regions; in Jeans. D. N. (cd.). Ausiraha. a i^eograpnv. Svdncy Univ. Press: 38-52. Johnson. 1). P (1982). — Sedimentary facies of an arid zone delta: Crascoync delta Western Australia, j. S'ed. Petrology. 52: 547-563. Johnson. D. W . (1919 ). — Shore proce.s.ses and shoreline development. Wiley, New York; Jutson. J. T. (1950). — The physiography (geomorphology) of Western Australia (3rd ed ). H V-v/ In.w. Geoi Sur. Bull. 95. King. C. A- M. (1966). — Icchniques in geomorphologs'. Arnold. London; 342. King. C'. M. ( 1972 ). — Beaches and Coasts (2nd ed.). Edward Arnold Ltd, London; 57()p. King, C. A. M. ( 1 975). — Technupies in geomorphology. Edward Arnold Lid. London; 342p- Langford-Smilh. E.. and Thom. B. O. (1969). — New South Wales Coastal Moiphology. J. Geoi Soc. Ausir. 16: 572-580. Liddlc, H. G.. and Scott. R., 1925-1940: The Greek English Le.xicon. Revised by H. S. Jones (9ih ed.). Oxford. Linton. D. L. (1951). — The delimitation of morphological regions: in Stamp. L. D. and Wooldridge. S W. (eds). London essays in geography. Longman. London. 199-217. Logan. B. W.. and Brown, R. C>. (1976). — Quaternary sediments, sedinieniary processes and environments; in Logan. B. W, et ai, Carbonate sediments of the west coast of Western Australia. 25th hit. Geoi Congi E.xcuroon tiuide .Vo. 4-75. Logan, B. \S .. and Cebulski, D. £► (1970). — Sedimentary environments of Shark Bay. Westem Australia, in Logan. B. W'. et uL. Carbonate sedimentation and Environments. Shark Bay, Western Australia. .Am. . l.v.vor. Peiroleum Geoi Mem . 13 : 1-37 Mabbuil. .1 A- (1968). — Review of Concepts of Land Classification: in Stewart. G. A. (ed.). Lund Evaluation. Macmillan of Australia, Melbourne: 1 1-28 McArthur. \V'. M. and Bettenay. E. (I960). — The development and dislnbulion of the soils of the Swan Coastal Plain, Western Australia. Soil PubI \o. 16. CSIRf). Melbourne. McCuIlagh. P (1978 ). — Modern Concepts In Geomorphology. Oxford I'niv. Press. Oxford: 128p. McKee. E. D. (ed ) (1979). — A study of global sand seas. VS Geo! Survey Prof Pap . M)?s2:Al%. Mitchell. C. ( 1973). — Terrain evaluation. Longman. London; 221p. Morgan, .1. P. (1967). Depositional processes and products in the deltaic environment: in Morgan. J. I. (ed.). Deltaic sedimentation: modem and ancient- Soc. Econ. Paleont. and Mineralogi.sf Spec. Pubi.. 15: 31- 47. Odum. F. P. (1971 ). — Fundamentals of ecology (3rd ed.). Saunders. Philadelphia: 574p. I’crnn. R. M. S. and Mitchell. ('. W. (1969). — An appraisal of physiographic units for predicting site conditions in and areas. .Mil. Eng. Exp. I-Aiab. Kept. 1111 . Vol. 1. Phleger, F. B. ( 1977). — Soils of marine marshes; in C'hapman. V. J. (cd.). Ect>s\'::lems o/ the World, h II i7 ciyaslaf m».vv-v/<7ii-v. Elsevier, Amsterdam 69-77. Purser. K. H. and Evans. G. (1973). — Regional sedimentation along the Trucial coast. S. E. Persian Gulf; in Purser, B. H. (Ed.), The Persian Gulf Springer-Verlag. Berlin; 21 1-231. 78 Journal of ihe Royal Society of Western Australia, Vol. 68. Part 3, 1986 Price. W. A. (1 955).— C'orrelaiion of shoreline types with offshore bottom conditions. Dept Oceanogr. Proj. 63. Texas A and M University- Read. J. F. (1974). — Carbonate bank and wave-built platform sedimentation. Edel Province, Shark Bay. Western Australia; in Logan, B. W. et al.. Evolution and Diagencsis of Ouaiernary Carbonate Sequences, Shark Bay. Western Australia. Am. .Iwrx’ Fclroleum GeoL. Mem. 22: 1-60. Russell. R. J. (1967). — River plains and sea coasts. Univ. of California Press; 84-8. Russell. R. J. and McIntyre. W. G. (1966). — Auslraltan tidal flats. Coastal Studies Scries No. 1 3. Louisiana Slate University I’ress; 48p. Semeniuk, V. (1980). — Mangrove conation along an eroding coastline in King Sound, Norih-wcsicm Australia. ./ tcol.. 68: 789-812. Semeniuk. V, ( 1 98 la). — Sedimenlology and the stratigraphic sequence of a tropical tidal Hat. northwestern Australia. Sedimentary’ GeoL. 29: 195- 221 . Semeniuk. V. (1981b). — Long-term erosion of the tidal flats King Sound. Northwestern Australia. Xlanne Geol.. 43; 21-48. Semeniuk. V (1983). — Mangrove distribution in northwestern Australia in relationship to regional and local freshwater .seepage. I cf’ctatio. 53: 1 1- Semeniuk, V. (1985). — Development of mangrove habitats along na coasts in north and northwestern .Australia, 1 egetatto. 60; 3-23. Semeniuk, V.. Chalmer, P. N. and LeProvost. L (1982). — The marine environments of the Dampier Archipelago. J. Row Stk: H es?. 65; 97-114. Shepard. P. F. (1963 ). — Suhnanne Geohgv (2nd ed.). Harper and Row. New York; 557p. Stephenson, T. A., and Stephenson, A. ( 1 972). — Life between tidemarks on rocky shores. Freeman and Co.. San Francisco; 425p. van Straaten. L. M. J. U. (1954). — C'omposilion and structure of Recent marine sediments in the Netherlands. Leid.se Geol. Mededel.. 19; 1-110. Thom. B. G. (1967). — Mangrove ecology and deltaic geomorphology: Tabasco. Mexico. ./. licoi 55: 301-343. Thom, B. G . Wright. L. D. and Coleman. J. M.{1975). — Mangrove ecology and dcliaic-estuarine geomorphology; Cambridge Gulf-Ord River, Western Australia. J. Eeol.. 63: 203-232. Thompson, R. W. (1968). — Tidal flat sedimcnialion on the Colorado River delta. Northwestern gulf of California. Geol. Soc. Am. Mem.. 107: I33p, Trewanha, G. T., Robin.son. F. H. and Hammond. E. H. (1968). — Fundamentals of physical geography. McGraw-Hill, New York: 384p. Tricart, J. (1972), — Landforms of the humid tropics, forests and savannas. Longman, London: 366p- Tumcr, F. J. and Weiss, L. E. (1963). — Structural analysis of meiamorphic tectnnites. McGraw-Hill, New York; 545p. Twidalc. C. R. (1968). — Geomorphology. Nelson. Melbourne: 406p. Valentin. H. (1952). — Die Kiisten dcr Erde. Petermanas Geog. Milt. Erganzungshefi: 246p- Wolff. W J. (cd.) (1983), — Ecology of the yi'adden Sea. A. .A. Balkcma, Rotterdam. Woods, P. J.. and Brown. R. G. (1975). — Carbonate sedimentation in an arid zone tidal Hat, Nilemah Embaymeni. Shark Bay, Western Australia: in Ginsberg. R. N. (cd.). Tidal Deposits. Springer-Veriag. Berlin: 223-232. Woolnough. W. G. (1920). — fhe physiographic elements of the Swan Coastal Plain. J. Roy. Sac. H esv. . Utstr.. 5: 15-19. Wright, I (1978). — River deltas; in Davis. R. A. (Ed ). Coastal Sedimentary Environments. Springer-Veriag. New York: 5-68. Wright. L. D.. Coleman. J. M.. and Thom. B. G., (1973) Geomorphic coastal variability, northwestern Australia. Coast. Stud. Bull.. 7: 35-64. Yonge. C. M. (1966). — The sea shore. Collins, London; 3 1 1 p. Zenkovich, V. P. (1967). — Processes of coastal development. Wiley Iniersciencc. New York: 738p. 79 Journal of the Royal Society of Western Australia, Vol. 68, Part 4, 1 986, p. 81-93. Australites from Hampton Hill Station, Western Australia by W. H. Cleverly W.A. School of Mines, Kalgoorlie, Western Australia, 6430 Manuscript received 19 February' 1985; accepted II February 1986 Abstract The largest known concentration of australites (Australian tektites) was on Hampton Hill pastoral station. Western Australia. Nearly 22 000 of them have been classified morphologically according to a “binomial” scheme based upon two views of the australite when oriented in ablation flight: the classification scheme is described in detail. Comparisons have been made with some other localized occurrences but their usefulness is limited by differences in definitions and bias in samples. Minor studies were made of the distinction between oval and boat shapes, the derivation of “conical” cores, and weight distribution amongst whole australites. Introduction This paper is an account of the australites collected on Hampton Hill Station. The western boundary of the station is 9 km east of Kalgoorlie and the homestead is a further 18 km east (Fig. 1). The property is owned and operated by Mr C. B. C. Jones and family. It includes the abandoned gold-mining centre of Bulong, once a considerable town — indeed, a municipality — and the sites of the former small mining townships of Kurnalpi, Boorara and Golden Ridge. It includes also the complex lake system comprising Lake Yindarlgooda, its north- easterly extension sometimes known as Lake Lapage, and the yet further extension shown on some maps as Cooragoogginc Lake. Thus the australites attributed variously to Bulong, Kurnalpi, Lake Yindarlgooda and Lake Lapage qualify for consideration here. There are well over 22 000 australites from Hampton Hill Station in collections. The principal units examined were the Hampton Hill components of the various Jones family collections totalling 14 155 and of the Tillotson family collections totalling 7 478. Small numbers in the British Museum (Natural History), National Museum of Victoria, South Australian Museum. Western Australian Museum and W.A. School of Mines totalling 294 were also examined. A few hundred others, principally in the Smithsonian Institution, American Museum of Natural History and the private collection of Mr P. J. Simmonds, but also as small representations in a number of official and private collections, were excluded from this study for various reasons. It is unlikely that the excluded specimens, which constitute about 2% of those known to be in collections, would affect significantly the results obtained. Hampton Hill Station contained the largest known concentration of australites. The only other area which has yielded a comparable number of specimens is that spanning pan of the South Australia-Nonhem Territory border. From that area have come the Kennett collection (7 1 84) and the Finke collection (1811), both of which are in the South Australian Museum. The Finke collection is a selection from a large number visually estimated by Ms J. M. Scrymgour as 10 000-12 000. Thus a total number in the order of 17 000 to 19 000 (exclusive of minor representations) is known to have been collected from that region, but only about half of them are in collections which arc available for study. The total number is comparable with that from Hampton Hill Station but there is a major difference in the areas involved. It can be estimated from the statements of Mr Kenneii quoted by Fenner (1940: 307) and the statements of Mr McTavish of the former Apatula Mission at Finke on the occasion of sale of australites to the South Australian Museum (J. M. Scrymgour pers. comm.) that the adjoining and partly overlapping provenances of the Kennett and Finke collections have a total area in the order of 22 000 km^. Baker (1956:65) gave closely the same figure for the provenance of the Kennett collection alone. Hampton Hill Station has little more than one seventh of that area; moreover, though australites were widely distributed, most located specimens were found on only a fraction of the station. The Hampton Hill australites would have been in the order of 20-30 times more concentrated than those of the South Australia- Northern Territory border region. Australites in the Tillotson collections are specifically located and a “centre of occurrence" could be calculated by w'cighling the geographical co-ordinates of each item in the collection by the number of specimens in the item (Cleverly 1976: 223). Additional collecting since 1976 has made it desirable to recalculate the “centre of occurrence" as 121"56.3'E.. 30‘’39.2'S.. a slight modification of the earlier result. The centre is w'iihin a broad, boomerang-shaped area marginal to the lakes (Fig. 1) within which more than 97% of the Tillotson specimens were found. The bell extends from a point south-east of Bulong clockwise through Taurus. Lake Penny, northern Lake Yindarlgooda. Jubilee and Lake Lapage, to the vicinity of Kurnalpi and occupies closely one quarter of the area of the station. Most of the specifically located specimens of the small collections 81 46961-1 Journal of ihe Royal Society of Western Australia, Vol. 68. Part 4. 1986 Figure 1. Hampton Hill Station (firm line boundarv) showing the boomerang-shaped area around the lakes within which more than 13 000 australiies were tound. which were examined are also from within the boomerang-shaped area. Nearly all the 5 339 specimens in the collection of Mr and Mrs B. C. Jones were found within the same area as a highly concentrated occurrence a few hundred metres diameter at approximately I21‘’57'E., 30"4rs. This real feature of the shower was near the edge of the lake in the general vicinity of Rocky Dam and only about 3km distant from the calculated “centre of occurrence". It was almost certainly identical with the small area known later as “Moriarty’s Patch", from which a Kalgoorlie lapidary and mineral dealer obtained many australiies. The Tilloison collections also contain 507 from the vicinity of Rocky Dam and Monarty’s Patch but the best represented area in their collections is somewhat further south-west in the general vicinity of Taurus, from which thev obtained nearly 3 000 specimens. The boomerang shape of the concentration may have originated in this manner. The australite shower was very uneven in density (Fenner 1935, McColl and Williams 1970). One patch of greater than average abundance of radius a few tens of kilometres was centred round the small area of extreme abundance collected by Mr and Mrs B. C. Jones. Subsequent modification of the shape depended on peculiarities common to many lakes in this part of interior Western Australia. The lakes lend to have rocky western and northern shorelines of relatively simple outline and highly complicated eastern and southern shorelines where there is a complex of gypseous dunes, innumerable small basins, inlets and billabongs. The lake basins arc visualized as migrating as the result of active pediplanalion on the first type of shoreline and aggradation on the other. On the north- western side, small intermittent streams such as Magnesite Creek brought australiies down towards the lake and onto the numerous flat alluvial fans which extend out over the exposed or thinly concealed rock pediment. Because the shoreline is convex to the north- west, the zone of concentration is similarly cur\’ed. On the south-eastern side, australiies were buried under the encroaching dunes. The western and northern shores have been the more readily accessible to collectors until four-wheel drive vehicles became generally available in 82 journal oflhc Roval Societv orWesicrn Australia. Vol. 68. I^ari 4. 1986 recent years. This circumstance may have exaggerated the natural difference in australile abundance between the north-western and south-eastern sides. The only located items of any size from outside the boomerang-shaped concentration are 196 australiles in the Tillotson collections from the east side of the extreme south end of Lake Yindarigooda and 461 in the J.L.C. Jones collection from “eastern" Lake Yindarigooda. Both items include some of the best examples of australiles with V-grooves (Fig. 6) discussed in the last section of this paper. The precise locality of the J.L.C. Jones material — suspected to be close to or identical with the Tillotson locality — was not disclosed. Thus the area of concentration as shown in Fig. 1 should perhaps be modified or extended slightly to include the southern tongue of Lake Yindarigooda. The remaining nearly 9 000 specimens in the Jones collections are unlocated except insofar as they are from Hampton Hill Station, but it is known that many of them came from the lakes, if not from the area of concentration. The proportion of spurious specimens delected was rather higher than usually found in collections. Most are small, worn, almost black fragments of uUramafic rock and are visually deceptive except that their thin edges show^ a distinctive green (serpentine) colour in transmitted light. Their source is a broad band of ultramafic rock forming the western margin of Lake Yindarigooda. Of the 22 087 specimens examined, 160 (0.7%) were spurious, so that the genuine australiles numbered 2 1 927. Morphological classification. A “binomial" system of classification has been used (Table 1) based upon two views of the australile whilst it is in flight orientation with the line of flight imagined to be vertically downward. The system allows for 46 shape types. Two additional typc.s — “conical cores'* and “aberrant forms" are recognised. The total of 48 types is more than in any classification except that of Fenner (1934) but is believed to be appropriate for a group of australiles three limes as numerous as any previously reported upon in this way. The classification is largely genetic. A brief statement of the author's understanding of the development of australile shapes therefore follows, but it is not necessary that the reader subscribe to these beliefs in order to use the system. In consequence of some major impact event, a rebound jet of melt was dispersed as millions of small masses with high velocity. These masses assumed shapes approximating to the equilibrium shapes appropriate to their rates of rotation and retained those shapes during cooling to form the primary bodies. Subsequently, the resulting cold solid bodies of glass encountered the earth’s atmosphere and because of a combination of very high velocity, high symmetry of the shapes and downward direction of flight, most of them adopted stable orientations relative to their flight paths; some had slight axial wobbles or other instabilities. Frictional melting with ablation stripping of the melt and other aerodynamic (secondary) processes modified the shapes of the primary' bodies to form the secondary bodies w'hich were decelerated to the very modest velocities of free fall. Some secondary bodies may have been broken on landing but weathering and erosion processes during terrestrial residence have been much more important contributors to their present shapes. The two shape factors used in classification will now be considered. (a) Shape seen when looking along the flight path. Though conventionally referred to as the “shape" it is only the profile which is used. It could thus be either the rear (posterior) or the front (anterior) view. If looking downward at the australile in vertically downward flight, the rear view is the plan view or “plan aspect" of Baker and Cappadona (1972). The shape is closely dependent upon the form of the primaiw body and the orientation adopted relative to the line of flight when it encountered the atmosphere. The author's views of those two things are as follows: — Spheres The form adopted by non-rotating masses of melt was the sphere. This was the commonest of all primary shapes. Orientation in ablation flight was decided initially by the chance position at atmospheric encounter, but as soon as some ablation stripping had occurred from frontal surface, forces came into play ensuring the subsequent stable orientation relative to the flight path (Chapman et ai 1 962: 1 6 et seq.). Oblate spheroids Slowly rotating masses adopted a shape approximating an oblate spheroid with the short axis as the axis of rotation. They oriented in ablation flight with the short axis parallel to the flight path. Certain other orientations which are sometimes figured would have the aerodynamic centre in front of the centre of gravity and would not therefore be stable. The rotation of these and other bodies was usually damped out during atmospheric encounter or was reduced to no more than a slight axial wobble. All sections normal to the flight path were circular. The protected posterior surface of ablation flight, the only remnant from which the form of the primary body can be judged, is only occasionally sufiiciently well preserved or large enough for a distinction to be made from the sphere (Chapman and Larson 1963: 4334. Oeverly 1974; 69). Prolate spheroids More rapidly rotating masses of melt became elongated normal to the axis of rotation, assuming shapes approximating to prolate spheroids or, more correctly, iriaxial ellipsoids, though it is only in the largest cores or exceptionally well preserved specimens that departure from circularity in sections normal to the length can be delected in the remaining short arc of primary surface (Cleverly 1979a). They oriented with the longest axis (or two longer axes) normal to the flight path and the shortest one parallel to it. The principal section normal to the flight path was either bluntly oval or more narrowly oval or parallel-sided with rounded ends in the sequence of higher rates of rotation and decreasing numbers of bodies. In some instances, melt flowed unequally towards the ends and the resulting body therefore tapered towards one end. Dumbbells In yet more rapidly rotating masses, migration of melt towards the ends was sufficient to cause development of a waist. These bodies oriented with the longest axis normal to the flight path and the shortest one {i.e. axis of rotation) parallel to it. Sometimes the outward flow of melt was not evenly directed, resulting in gibbosities of unequal size. Such bodies are conventionally referred to as asymmetrical dumbbells though the asymmetry is usually only in respect of the ends, high symmetry being present relative to the other principal planes. The asymmetrical dumbbells oriented with the length otT the normal to the flight path, the heavier end inclined forward into the pressure at an angle dependent on the degree of inequality (Chapman e/ a/. 1962: 19). 83 Journal of ihc Royal Society of Western Australia, Vol. 68. Part 4. 1986 Apioids A few masses rotated sufficiently rapidly for the waists of dumbbells to thin to disappearance yielding two bodies approximating to apioid shape. As they were no longer held together in a rotating system, they presumably flew off tangentially. The largest ones with the longest liquid lives may have made progress towards blunter or even spherical shape. They oriented in ablation flight with the main body forward and the tail obliquely backward, the angle of obliquity depending upon the degree of tapering of the shape. Table 1 Shape names for ausfralifes* \ ► \ Rear \ (plan) \ view Side or\ y f end view (elevalion) \ o round o broad oval 5 L/w >1 O o narrow oval 2 ^ L/W >'*/3 ( ) CO co (asymmetrical ) dumbbell O boat L/W >2 O teardrop O'sO? flanged form button flanged broad oval flanged narrow oval flanged boat flanged (asymmetrical ) dumbbell flanged teardrop (cored) disc or plate ( cored ) disc ( cored ) broad oval plate { cored ) narrow oval plate boat- plate ( cored ) teardrop - plate (cored) ^ bowl ( cored ) round bowl (cored) broad oval bowl ( cored ) narrow oval bowl boat- bowl teardrop - bowl canoe form - broad oval canoe narrow oval canoe boat- canoe (asymmetrical) dumbbell- canoe - indicator type I round indicator I broad oval indicator I narrow oval indicator I boat - indicator I (asymmetrical ) dumbbell- indicator I teardrop - indicator I O lens form lens broad oval lens narrow oval lens boat- lens (asymmetrical ) dumbbell- lens teardrop - lens 0' different (Table 3) and the reasons not hard to Table 2 Morphological classification and weights of australites from Hampton Hill. Station, W. A. Shape type Numbers of specimens Weights of complete specimens Complete Broken Total Lightest g Heaviest g Mean g Button 1 1 2 1.95 Disc 1 — 1 0.16 Round bowl 5 15 20 0.19 0.42 0.30 Round indicator! 49 6 55 0.13 4.92 2.01 Lens 2 463 1 244 3 707 0.1 1 4.86 0.88 Round indicator II 8 1 9 0.81 22.38 4.54 Round core 1 584 467 2 051 0.47 71.29 5.33 Flanged broad oval 1 — 1 2.31 — — Broad oval bowl 7 6 13 0.24 0.53 0.35 Broad oval canoe 4 5 9 0.97 1.43 1.20 Broad oval indicator I 1 — 1 3.43 — — Broad oval lens 298 69 367 0.14 4.74 1,13 Broad ova! indicator 11 1 — 1 3.48 — — Broad oval core 408 64 472 0.53 101.12 7.42 Flanged narrow oval — 1 1 — — — Narrow ova! bowl 2 6 8 0.11 0.13 0.12 Narrow oval canoe 3 — 3 1.64 2.72 2.07 Narrow oval indicator I 1 — 1 0.92 — — Narrow oval lens 345 136 481 0.28 6.40 1.38 Narrow oval indicator II 2 — 2 2.31 4.15 3.23 Narrow oval core 289 86 375 0.93 69.25 6.74 Boat — canoe 1 2 3 2.85 — — Boat — lens 167 101 268 0.21 9.10 1.71 Boat — indicator II 6 — 6 3.61 6.10 4.80 Boat — core 142 67 209 1.06 42.00 8.30 Dumbbell — canoe 1 3 4 1.34 — — Dumbbell indicator I 3 1 4 1.12 6.00 4.23 Dumbbell — lens 249 388 637 0.26 8.16 2.08 Dumbbell — indicator II 4 3 7 3.37 6.30 4.84 Dumbbell — core 113 135 248 1.11 47.00 9.02 Teardrop — indicator I 1 — 1 2.85 — — Teardrop— lens 224 54 278 0.13 5.79 1.44 Teardrop — core 57 4 61 1.30 15.15 4.86 Conical core 1 393 18 1 411 0.27 17.94 2.78 Abberant forms 159 37 196 0.39 23.41 2.68 7 993 2 920 10913 Mean 3.08 Fragments 10 773 Flakes and flaked cores 241 Total 21927 87 Journal of the Royal Society of Western Australia, Vol. 68. Part 4, 1986 Table 3 General classification of australites from Hampton Hill Station compared with other localized samples General classification Hampton Hill Station, W.A. S.A. — N.T. border (Fenner 1940) Finke area (Cleverly unpub.) Number Percent Percent Percent Complete or essentially so 7 9931 M0913 2 920J 10773 241 36.5 1 > 49.8 13.3 -1 49.1 1.1 50.8 1 > 91.2 40.4 J 8.7 0.1 44.8 1 > 65.7 20.9 -I 32.2 2.1 Fragments Flakes and flaked cores Number of specimens..., 2! 927 3 920 1 811 find. As an example, the “classified” sub-class B2a (Fenner 1940: PI. VIII) consists of broken pieces of indicators II which might belong to any elongate group whatever: they would be unclassifiable in the more exacting system used here. The Finke collection from an adjoining area has been classified by the writer using the same system as for Hampton Hill, but this also shows considerable difference arising, at least partly, from the bias of the sample which is a chosen one sixth of the material offered for sale. The percentage of identifiable specimens is considerably inflated and the unidentifiable fraction correspondingly reduced. 2. A preliminary view of the state of weathering is possible from a consideration of the number of shape types present. Despite the large number of australites, only 35 of the 48 shape types are represented, the majority of them very poorly. Fourteen categories (6 lens form, 7 types of core and the aberrant group) account for 98.6% of the identifiable specimens. This reflects the severely weathered state of the material. These figures compare much more closely with the Finke collection which has 30 types represented, 14 of them accounting for 96.2% of identifiable specimens. The degree of weathering can be examined more conveniently by the method of item 4 below. 3. Figures may be extracted for plan or elevational view shaoes by adding the figures for individual items in the rows or columns of Table 1 . The plan view shapes (Table 4) reflect in a general way the proportions of the primary bodies. Spheres are not distinguished from oblate spheroids and the shape proportions have been influenced by later events e.g. the separation of dumbbells during ablation flight. Having in mind the differences in definition for items ranging from broad oval to dumbbells and also canoes, adequate comparision with collections classified by other systems is not possible. Even a collection from Lake Yindarlgooda (Chalmers et al. 1976: 18) with the small percentage of canoes re-distributed proportionally shows values widely different from those found here (Table 4). Some difference from the biased Finke collection is not unexpected. 4. The elevational view shapes exclusive of aberrant specimens (Table 5) reflect especially the effects of aerodynamic and terrestrial shaping processes. When weathering is well advanced, as at Hampton Hill, it is convenient to group the frail types viz. flanged forms, discs and plates, bowls and canoes as a single item. The percentages of these and of those australites still in progress via indicators to a stable end form is extremely small. Almost all medium-sized specimens have reached what is probably a final stable lens form (mean weight 1.02 g) and almost all larger specimens have lost stress shells as the result of aerodynamic or lerresmal processes to reach a final stable core form (mean weight 4.96 g). Lens forms and cores comprise 96.8% of the Hampton Hill australites. In spite of the bias in selecting the better preserved material for the Finke collection, the same general trend is clearly observable (Table 5). Table 4 Plan view shapes of australites from Hampton Hill Station and Finke. N.T. Hampton Hill Station, W.A. L. Yindarlgooda (Chalmers er fl/. 1976) Finke, N.T. (Cleverly, unpub.) Shape Number •Adjusted numbers Percent Percent Percent Round 5 845 7 074 66.0 50.5 71.8 Broad oval 864 1 046 9.8 21.6 8.8 Narrow oval 871 871 8.1 lO.l Boat 486 486 4.5 17.8 3.9 Dumbbell 900 900 8.4 7.4 4.6 Teardrop Conical core Aberrant 340 1 411 196 340 Re-distributed Eliminated 3.2 2.7 0.8 No. of specimens 10913 10717 10717 109 1 183 • Adjusted numbers and thence adjusted percentages are obtained by eliminating the aberrant forms and distributing the conical cores between round and broad oval groups in the proportions present in the original statement. 88 Journal of the Royal Society of Western Australia, Vol. 68, Part 4, 1986 Table 5 Elevational shapes of australites from Hampton Hill Station, W.A. and Finke, N.T. Elevational shape Hampton Hill Station, W.A. •Finke, N.T. Number Percent Percent Flanged, disc and plate, bowl and canoe forms 65 0.6 2.5 Indicator 1 62 0.6 8.3 Lens forms 5 738 53.6 49.4 Indicator II 25 0.2 3.1 Cores, including conical 4 827 45.0 36.7 Number of australites to 717 1 183 •Cleverly, unpublished study Lens forms and cores The mean weights of the Hampton Hill lens forms increase with increasing elongation from round to dumbbell; the teardrop-lens also conforms if it is conceded that it formed from half the primary body and the mean weight should therefore be doubled. The cores, with one exception also show an increase (Table 6). The trend is present in the lenses of large samples from Earaheedy Station and Finke, N. T. but irregularly and imperfectly. The regularity of increase is best in the largest samples. The reason for the increase is unknown but it is advanced most tentatively that the reason is related to the progressive change in the ratio of cross- sectional area to volume (and mass) as elongation increases. This ratio could influence such variables as deceleration, heating and the expansion and contraction of the anterior shell. Table 6 Mean weights of lens forms and cores Shape type Lens forms Hampton Hill Station. W.A. g Cores Hampton Hill Statiop, W.A. g Lens forms •Finke, N.T. g Lens forms •Earaheedy Station, W.A. g Round 0.88 5.33 1.78 1.30 Broad oval 1.13 7.42 1.99 1.57 Narrow oval 1.38 6.74 2.53 1.50 Boat 1.71 8.30 3.10 2.35 Dumbbell 2.08 9.02 2.66 2.11 Teardrop (x 2) 2.88 9.72 5.32 3.94 No. of specimens 3 746 3 986 347 259 •Cleverly, unpublished studies Total and average weights Total weight of 7 993 complete or essentially complete Hampton Hill australites Weight of 1 3 934 other specimens Total weight of 21 927 australites 24616g 18 934 g 43 550g The mean weights calculated from the above are shown in Table 7. Such figures could be greatly influenced by the degree of care in collecting. Comparison is made in Table 7. with Myrtle Springs Station where australites are conspicuous against their background and the collecting was carefully done by a Museum party, the Port Campbell area, which was repeatedly carefully collected, and the Florieton area, a first amateur collection effort. Some authors omit either the mean weignt of whole specimens or the mean weight of all specimens. The second of these, if small, is a guide to the reliability of the first. Without that reassurance it must be suspected that high mean weights could be artifacts of careless collecting or some other cause, such as the transport of australites (Cleverly 1976: 23 1). Table 7 Mean weights of auslraliies from Hampton Hill Station and other localities Hampton Hill Stn., W.A. Myrtle Springs (Corbett 1967) Port Campbell (Baker, 1956) "Florieton, S.A. (Mawson 1958) Number of complete australites 7 993 212 812 Mean weight of complete australites (g) 3.08 2.73 3.78 Total number of australites including fragmenLs and flakes 21 927 175 573 1 475 Mean weight of all specimens (g) 1.99 1.25 ••1.45 2.85 •Mean weights calculated from data of Mawson (1958) ••Recalculated from data of Baker ( 1 956) Weight distribution The Tillolson collections of located specimens are numerically large and have resulted from careful searching. They therefore provide a rare opportunity to study weight distribution in a large sample of whole australites. Forty three trays of australites provided from time to time by Mr R. G. Tillotson were accepted as a sample of the stack of trays. They contained 6165 australites of which 2157 were whole or essentially so. .Alt whole specimens were weighed, the frequency of the weights in 0.01 g intervals plotted as a histogram and from it a curve drawn and smoothed slightly (Fig. 4). Figure 4. — Frequency of weights for 2 103 whole australites out of sample of 2 1 57. The balance of 54 had weights 1 0.0 1 -58.25 g lying beyond the upper weight limit of the figure. 89 Journal oflhc Royal Society of Western Australia. Vol. 68. Part 4. 1986 The mode is closely 0.9 g. Above that weight the frequency declines at first rapidly and then more slowly. There are 2103 specimens under the cur\'e in Fig. 4. The remaining 54 specimens arc dccrcasingly numerous between 10.01 and 58.25 g. /.e. over a width nearly five times that of the figure. This portion of the curve may reflect the primary event which scattered small molten masses having decreasing chances of holding together with large size. If so. the distribution has survived — at least in a general way — the rigors of reduction in size and re-shaping by aerodynamic and terrestrial processes. Several causes may have contributed to the form of the curve to the left of the mode. A minimum size of primary^ body is presumably needed if it is to survive atmospheric entry and leave a detectable secondary- body. Weathering and erosion processes have reduced the sizes of most australites and perhaps destroyed some of the smallest ones. But the major difficulty is in observing small specimens in the field even for the most painstaking collectors. A common difficulty, that persons collecting for reward may ignore small specimens and flakes as being useless to lapidaries and therefore valueless does not apply to the collections used here. The curve may be compared with that of Fenner (1934. Fig. 4). whose approach was radically different. Fenner used the average weights of those morpholigical groups having averages less than 3 g plotted in 0.2 g intervals to obtain a very irregular bimodal frequency distribution of 1858 specimens, and from that distribution a curve was drawn. Some of the groups may have included individuals with weights exceeding 3 g: on the other hand, excluded groups may have contained individuals with weights less than 3 g. The majority of specimens came from Israelite Bay (Fenner 1934; 65) which is about 370 km south-south-east of Hampton Hill Station. Despite the different approaches and localities, the resemblances between the curves are close. Both show modes ofc. 0. 8-0.9 g, a strongly concave frequency curve asymtolic to the weight axis for the higher weights and a slightly convex curve for weights below the mode i.e. plunging increasingly steeply as size decreases and detection becomes increasingly difficult. Chemical composition and specific gravity The analysis of an australilc from Kurnalpi was given by Taylor (1962) and of a second from the same locality by Taylor and Sachs (1964): trace element data were included. Analyses of a specimen from each of Lake Yindarigooda and Lake Lapage were published by Chapman and Scheiber (1939) and allotted to the "normal" chemical type. The specific gravity of australites from the same two localities was studied by Chapman (1971, Figs. 4(d) and 5(b)) in samples of 912 from Lake Yindarigooda and 1094 from Lake Lapage. The frequency diagrams show single, strongly pronounced modes (70%-80% of samples) in the 2.45- 2.46 interval. Mason (1979) studied a sample of 61 australites from Lake Yindarigooda and found the specific gravities of 93% of them in the 2.45-2.46 interval: his three accompanying chemical analyses represent almost the whole range of specific gravity, or inversely, the range of silica content. Individual specimens Details have been published elsewhere of an exceptionally large broad oval core weighing 101.1 g found between Kumalpi and Jubilee (Cleverly 1974), several small bowls (Cleverly 1979), a variety of aberrant forms (Cleverly 1982a) and hollow australites (Cleverly 1982b). A further variety of forms is shown in Fig. 5. They include two australites in the Tillotson collections found as weathered, well separated fragments which could be refitted together (Figs. 5.32 and 5.33). Refitting of fragments usually requires both careful search and good documentation before the possibility of reunion is recognised. Both circumstances are lacking for many Western Australian collections. The most striking or the australites from Hampton Hill Station are those having patterns of V-grooves upon posterior surfaces, and occasionally, on anterior surfaces also (Figs. 5.35-39 and Fig. 6). Individual grooves are usually several millimetres long and not more than a millimetre deep. They may bifurcate but such grooves are seldom abundant (Fig. 6.1): no specimen shows a perfection of development like the core ftom Hattah, Vic. described by Baker (1973; 205 and PI. 28), though a closely comparable core is known from Western Australia (Simpson 1902: PI. I). The grooves are especially common on larger cores but occur also on some lens and aberrant forms. Their pattern may be radial on round and broad oval cores and then increasingly longitudinal on the more elongated forms. The centre of radiation is most commonly the posterior pole but can be an eccentric feature such as a bubble cavity (Fig. 6.39). When a core has been broken early in its terrestrial history, grooves may radiate from a point within the remnant piece; a particularly fine e.\ample from elsewhere has been illustrated by Cleverly and Sciymgour (1979: Fig. 2.10). Grooves may be sufficiently numerous to form “bird track” or reticulated patterns, especially on round and broad oval cores, or on the other hand, so uncommon as to appear unsystematic, sometimes even a single groove. l igure 5.— 4uslraliles from Hampton Hil) Slalion. Western Australia, natural size unless othervvise slated. In elevational views, direction of flight is towards bottom of page .Abbreviations used are;— p.s. posterior surface, a.s. anterior surface and s.e. side elevation. I Button, p.s, with thin travertine coating. Round indicator 1. p.s. and below it. elevation seen through gap in flange. 3. Indicator 1 of’small" button, p.s.. x 2. 4 Round indicator I p.s. and s.e. bclow^5. Lens, cievaiioti. \ 1 .5 showing pale "flange band” posterior to rim. 6. Incomplete detached round flange, p.s.. x 2. 7, Round indicator II. p.s.. X - and s.e below X. Round conical cure. s.e.. x 2. 9. "Sniair’ mund core. s.e. 10 •’Small*' round core. s.e. x 1.25 II Ranged broad oval, p.s with side and end cUnatums below. 12. Broad oval bowl. p.s.. x 2. 13. Broad oval canoe with "tortoise-sheir posterior surface. -S.e.. 2 . 14. “Small” ** *' ' 1 ^ Wedged liroad oval core. a.s. showing wedge, x 1.5 16. Narrow oval lens. a.s. showing flow ridges. 1 7. Narrow oval lens with butt ot flange, p.s. J8. Boat indicator 11. p.s. above and a s below retaining small areas of sta-ss shell at each end. 19. Wedged stout-waisted dumbbell core, s.e above, a.y below. 2 photographs used in Figs. 5 and 6 was done by Mr M.K. Quartermaine. MsJ.M. Weame drafted Table ) and Figs 1-4 References Baker, G. (1956). — Nirranda sirewnfield australites, south-east of Warmambool, Western Victoria. Mem natl. Mus. Viet.. No. 20; 59-172, Baker, G. (1972). — Largest australite from Victoria, Australia. Mem. natl. Mus. Viet: No. 33; 125-130. Baker. G. ( 1 973). — Australites from the Murray-Darling confluence region, .Australia. Mem. natl. Mus. Viet.. No. 34; 199-207. Baker. G. and Cappadona. W. J. (1972). — Smallest known complete australite. Mem. natl. Mus. Viet.. No. 33; I3I-135. Cappadona. W. J. (1981). — Notes on a fragment core australite. Proc. Roy. Soc. Vtcl.. 92; 207-208. Chalmers. R. O., Henderson. E. P. and Mason. B. (197,6). — Occurrence, distribution and age of .Australian tektites. Smithsonian contributions to the earth sciences. No. 17. Chapman. D. R. (1964). — On the unity and origin of the .Australasian tektites. Geochim. et Cosmochim. .icta. 28: 841-880. Chapman. D. R. (1971). — Australasian tektiie geographic pattern, crater and ray of origin, and theory of tektite events. J. Geophys. Res.. 76: 6309-6338. Chapman, D. R. and Larson. H. K. ( 1 963). — On the lunar origin of tektites. J. Geophys. Res . 68: 4305-4358. Chapman, D. R., Larson. H. K. and Anderson. L. A. (1962). — Aerodynamic evidence pertaining to the entiy of tektites into the earth's atmosphere. N.A.S.A. Technical Report R- 1 34. Chapman, D. R., Larson. H. K. and Scheiber, L. C. (1964). — Population polygons of tektite specific gravity for various localities in Australasia. Geochim. et Cosmochim. Acta. 28; 821-839. Chapman. D. R. and Scheiber. L. C. (1969). — Chemical investigation of Australasian tektites. J. Geophys. Res.. 74; 6737-6776. Cleverly, W. H. (1974), — Australites of mass greater than 100 grams from Western -Australia. J. Roy. S(K V'est- 57; 68-80. Cleverly, W. H, ( 1976). — Some aspects of australite distribution pattern in Western Australia. Rcc. IVe.it. Ausf. .\fus.. 4: 2 1 7-239. Cleverly. W. H. (1979a). — Broad oval australite core from Muntadgin, Western Australia. Rec. V'esi. Ausi Mus.. 7: 245-253. Cleverly, W. H. (1979b). — Morphology of small australites from the Eastern Goldfields, Western Australia. J. Rov. Soc. iVesl. Aust.. 61: 119-130. Cleverly. W. H. (1982a). — Some aberrant australite forms from Western Australia. J. Roy. Soc. Vest. Aust . 65: 17-24. Cleverly. W. H. (1982b). — Hollow australites from Western .Australia. Rec. iVest. -Jws/- Mus.. 9: 361-369. Cleverly. W. H. and Scrymgour, June M. (1978). — Australites of mass greater ihan 100 grams from South .Australia and adjoining states. Rec. S. .Aust. Mus.. 17: 321-330. Corbett, D W. P. (1967). — Australites from Myrtle Springs Station. South Australia. Rec. S. Aust. Mus.. 15: 561-574, Fenner, C. (1934). — Australites, Part 1. Classification of the W. H. C. Shaw collection. Trans. Rov Soc-S 4ust.. 58; 62-79. Fenner, C. (1935). — Australites, Part II. Numbers, forms, distribution and origin. Trans. Roy. 9of. S. Aust.. 59: 125-140. Fenner, C. (1938). — Australites, Pan III. .A contribution to the problem of the origin of australites. Trans. Roy. Soc. S.Ausi.. 62: 192-216. Fenner. C. (1940). — Australites. Part IV. The John Kenneit collection with notes on Darwin Glass, bediasites etc. I'rans. Roy. Soc. S. .4ust.. 64; 305-324. McColl, D. H. and Williams. G. E. (1970). — Australite distribution pattern in southern central Australia. Nature. 226: 1 54-155. Mason. B. (1979), — Chemical variation among .Australian tektites. fn R. F. Fudali. editor. Mineral Sciences Investigations 1975-1977. Smithsonian Contributions to the Earth Sciences. No. 22: 1 4-26. Mawson. D. (1958). — Australites in the vicinity of Florieton. South Australia. Trans. Roy. Soc. S. 4i^/.. 81: 161-163. Simpson. E. S. (1902). — Obsidianites. In Notes from the departmental laboratory. Geol. Surv. West. Aust. Bull . No. 6: 79-85. Taylor. S. R. (1962). — The chemical composition of australites. Geochim. et Cosmoentm. Acta. 26: 685-722. Taylor. S. R. and Sachs. M. (1964). — Geochemical evidence for the origin of australites. Geochtm. et Cosmochim. .Acta. 28: 235-264. Figure 6. — Australites from Hampton Hill Station. Western .Australia. Except where otherwise stated, all views are posterior surfaces of cores showing V- grooves and are natural size. 1-2. Broad oval. 3. Side elevation of broad oval showing extension of groove system to anterior (lower) surface. 4. Narrow oval, upper view of posterior surface, lower of anterior surface. 5-7. Narrow ovals. 8-10. Boats. II. Tapered boat. 12. Siout-waistcd dumbbell. 13, Dumbbell. 14. Asymmetrical dumbbell. 1 5. Slightly assymmeirical. sloul-waisied dumbbell. 16. Teardrop. 1 7-1 8. Conical cores. 19. Narrow oval lens, upper view the supposed anterior surface, lower the posterior surface. 20. Fragment of nut-like aberrant form, anterior surface above, posterior below. 2 1 . The two major surfaces of a canoe-like aberrant form, flight orientation indeterminate. 93 Journal of the Royal Society of Western Australia, Vol. 68, Part 4, 1986, p. 95-103. High-temperature retrograde adjustments in some Precambrian granulite-facies rocks at Albany, Western Australia by N. C. N. Stephenson Department of Geology and Geophysics, University of New England, Armidale, N.S.W. 2351, Australia Manuscript received 17 September 1985: accepted 15 April 1986 Abstract Detailed electron microprobe study of Precambrian meta-igneous enderbitic gneiss and mafic granuiite from Albany, Western Australia, reveals evidence of compositional adjustments in the high-grade minerals during the early stages of post-melamorphic cooling. These two lithologies contain the assemblage plagioclase -i- orthopyroxene -i- clinopyroxene + hornblende -P K-felspar + quartz ilmenile + magnetite ± bioliie, allowing the application of several clement-distribution thermometers. Intra- and intergranular compositional heterogeneity in pyroxenes and opaque oxides appears to record a period of incomplete readjustment during cooling from 750-800”C to about 600-650“C. Hornblende growth is correlated with this period. Re-equilibration of feldspars apparently extended to a lower temperature, perhaps about 500“C. Quantitative reconstruction of metamorphic cooling histories in rocks displaying evidence of retrograde readjustments is made difficult by several factors. These include the limitations of ««lement-distribution thermometers, the possible failure of co-existing mineral pairs to achieve or preserve chemical equilibrium at any particular stage (or stages) during readjustment, and the difficulties in recognising any such re-equilibrated compositions that may exist. Introduction The homblende-granulite facies rocks occurring in the south-coast region of the Precambrian Albany-Fraser Province, Western Australia, show obvious petrographic evidence of low-T retrograde modification of variable intensity; e.g., incipient to pervasive development of low-T alteration products such as fibrous amphibole, chlorite, clinozoisite, sericite and pinite. In addition, detailed microprobe studies of certain gamet-biotite (- cordierite) and orlhopyroxene-clinopyroxene associations have revealed that these rocks have also experienced partial chemical readjustment, involving cation exchange or transfer between coexisting phases, during the early cooling stages of regional metamorphism; i.e., at relatively high temperatures, often still within the granuiite facies environment (Stephenson 1979, 1984). Because these higher-T readjustments are not evident during routine optical examination of thin sections, their geographic extent and the identities of all the minerals affected are not yet known. This paper investigates further microprobe evidence of these relatively high-T retrograde modifications, this time in meta-igneous gneiss and granuiite at Albany (Fig. 1). Geological setting The Albany-Fraser Province is a belt of Proterozoic high-grade gneisses and granitic plutons bordering the Archaean Yilgam Block. In the south-coast region of this Province the gneisses are predominantly granitic in composition, with occasional metasedimentary and metabasite bands and lenses. They belong mainly to the upper amphibolite facies (sillimanite-orthoclasc zone) but large domains, several kilometres across, of homblende-granulite facies rocks (characterised by plag -t- opx A cpx f hbe, especially in mafic lithologies) are scattered through the terrain. These granuiite domains show intermediate- to low-pressure characteristics according to the criteria of Green and Ringwood (1967). The genetic relation between the lower- and higher-grade rocks has not been established. The present study concerns a homblende-granulite domain outcropping in the township of Albany (Fig. 1). This domain includes a large unit of enderbitic gneiss which locally encloses numerous small fragments of mafic granuiite. These two lithologies were selected for detailed study because (i) they show minimal development of low-T alteration products which might obscure the more subtle high-T retrogression, (ii) they contain identical mineral assemblages (although relative abundances differ), and (iii) they may be assumed — because of their close spatial association — to have experienced identical regional-mctamorphic histories. The body of enderbitic gneiss measures roughly 2 x 1 km in outcrop area, and is centred on Mt Adelaide in the eastern part of Albany township (Fig. 1). Excellent exposures are provided by road cuttings and by a section of coastline at the eastern edge of the body. The rock is characterised by a dark green colour on iresh surfaces, typical of felsic rocks with charnockitic affinities. An igneous origin for the body is indicated by sharp, apparently intrusive contacts with adjacent gneisses near 95 Journal of the Royal Society of Western Australia. Vol. 68. Part 4. 1986 Figure 1 . — Geological sketch map of the eastern pan of Albany township. Western Australia, showing sample locations. Wooding Point, and by the presence of numerous inclusions of mafic and feisic granulite and gneiss which appear to be xenoliths. The chemical analyses listed in Table 1 indicate a granodioritic to tonalitic composition for the enderbitic gneiss, and a dioritic composition for the mafic xenoliths. High Fe/Mg ratios are a notable feature of both rocks. Foliation in the enderbitic gneiss is defined by small, streaky mafic aggregates. It strikes E-W and dips nearly vertically throughout the body. Locally (e.g., near Wooding Point), the mafic xenoliths are strongly flattened parallel to the foliation in the surrounding enderbitic gneiss (Fig. 2). Figure 2.— Mafic granulite inclusions flattened parallel to foliation in the enclosing enderbitic gneiss. 650 m SSW of Wooding Point, Albany. Petrography Enderbitic gneiss This rock comprises a medium-grained, granoblastic- polygonal aggregate of plagioclase, quartz and K- feldspar, and wispy mafic streaks dominated by orthopyroxene and hornblende, plus minor clinopyroxene, biotite and ilmenite. Rare accessories include titanomagnetite, allanite, apatite and zircon. The mode of the specimen selected for probe analysis (sample 103a) is listed in Table 1. The major minerals range in grain size from 0.5 to 5 mm, but are generally less than 2 mm. Equant xenoblasiic grains predominate, but some grains of feldspar and orthopyroxene are rectangular in shape, elongate parallel to (010) or the c- axis, respectively. Because rectangular grain shapes are more common in these minerals in igneous than mnulile fabrics, they may be relic outlines inherited from the original texture. K.-feIdspar and quartz occasionally tend to embay other minerals, especially plagioclase. These embayment textures are most strongly developed at the western edge of the enderbitic gneiss body where they result from pronounced K-Si metasomatism emanating from the adjacent Albany Adamellite (Stephenson 1974). The widespread occurrence of similar, though weakly developed, embayment textures in the enderbitic gneiss remote from the obvious metasomatised zone may suggest the occurrence of mild metasomatism throughout the unit. This question is addressed again below. Hornblende (X = light brown, Y = brown, Z = dark green) and minor biotite (X = straw, Y = Z = dark reddish brown) occur occasionally as discrete grains, but more commonly as overgrowihs on opaque oxide or pyroxene, and as poikiloblasis enclosing opaque oxide, pyroxene and plagioclase (Fig. 3). Thus there is good I evidence that the hornblende and biotite post-date the enclosed minerals and therefore may be retrograde in origin. Retrograde hornblende replacing pyroxene in , granulites is commonly acicular or fine-grained granular in form, and typically contains numerous quartz inclusions (e.g. Beach 1974, Sills 1983). In the enderbitic gneiss at Albany the hornblende does not show these features; instead it is medium-grained and rarely contains quartz inclusions. However, it is sometimes separated from adjacent pyroxene by a narrow film of quartz (Fig. 3A), which is interpreted here as genetically equivalent to the more usual included quartz generally regarded as a 'by-producf of pyroxene-hydration. QUARTERNARY I Alluvium & Soil I Aeolian Sand PRECAMBRIAN : : : : Albany Adamellite Enderbitic Gneiss ;;i j! Granitic Gneiss + Sample Locality PR/NCESS ROYAL King Point HARBOUR Wooding Point 96 Journal of Ihc Royal Society of Western Australia. Vol. 68. Part 4. 1986 Table 1 Chemical analyses and modes of enderbitic gneiss ( 1 03a) and mafic granulite (282) samples used for microprobe analysis Sample No. Catalogue No.* 103a 54579 282 R45938 Si02 61.81 54.61 TiO-y 0.80 1.12 AI2O3 17.25 17.77 ^^2^3 1.57 0.85 FeO 5.65 9.14 MnO 0.17 0.22 MgO 1.28 2.21 CaO 5.00 6.88 Na-)0 4.00 4.52 K26 2.33 1.60 P2O5 0.24 0.40 H2O 0.52 0.51 Total 100.62 99.83 Trace Elements (p.p.m.) Rb 48 37 Sr 401 470 Y 31 65 Zr 442 357 Ba 3950 # Modes (vol. %) Quartz 13 0.5 k-feldspar 10 8 Plagioclase 60 58 Orthopyroxene 8.5 15 Clinopyroxene. 0.5 0.5 Hornblende 4.5 16 Biotite 1.5 tr Others 2.0 2.0 • Catalogue numbers refer to the collections of the Department of Geology, University of Western Australia (54579), and the Department of Geology and Geophysics. University of New England (R45938). # Not determined Analytical methods: trace elements by XRF; major elements by XRF ( 1 03a) and wet chemistry (282). Analysts: N.C.N. Stephenson and G.i.Z. Kalocsai. Another possibility — that the hornblende and biotite overgrowths are relic igneous features — is perhaps less likely in view of the thorough textural reconstitution that prevails in the gneisses and granulites throughout the region. The most common low-T secondary material is very fine-grained, fibrous to flaky, pleochroic green phyllosilicale with birefringence ranging from low (chlorite?) to quite hi^ (biotile?). It occurs along grain boundaries and microfractures in orthopyroxene, feldspar and quartz, and is assumed to be related to low- T fluids. Mafic granulite inclusions This rock is a granoblastic-polygonal aggregate of plagioclase, hornblende and orthopyroxene, plus vanable amounts of clinopyroxene, biotite, quartz and K-feldspar. Minor accessories include ilmenite, titanomagnetite, apatite and zircon. Biotite, where present, shows a preferred alignment imparting a foliation obvious in thin section, if not in hand specimen. K-feldspar embays plagioclase only rarely. Hornblende (X = light brown, Y = brown. Z - dark green) occasionally encloses pyroxene, but more commonly occurs as discrete grains. The mode of the specimen selected for probe analysis (sample 282) is listed in Table 1. This specimen contains a lower content of mafic minerals than usual for this lithology. Mineral chemistry and thermometry One sample each of enderbitic gneiss (103a) and mafic granulite inclusion (282) have been selected for detailed electron microprobe study using a JEOL JSM-35 SEM with a Tracor-Northem TN 2000 energy dispersive system, employing the instrumental conditions and data reduction techniques outlined by Ware (1981). The results are reported and discussed below. Figure3. — Photomicrographs of enderbitic gneiss, sample 103a. Bar -0.3 mm. PPL. A. Hornblende (dark grey; oblique cleavage traces) mantling clinopyroxene (light grey) and opaque oxide (black). Note the thin film of quartz (white) along portion of the clinopyroxene-homblende boundary. B. Poikiloblastic hornblende (medium grey) enclosing orthopyroxene (dark grey), opaque oxide (black), plagioclase and quartz (both white). 97 Journal of ihc Royal Society of Western Australia. Vol. 68, Part 4, 1986 Pyroxenes The probe analyses in Table 2 show that the analysed pyroxenes are characterised by high Fe^+/Mg ratios reflecting the host rock compositions (Table 1). The A1 contents are low compared with many granulite-facies pyroxenes, but typical of those from relatively low- pressure granulite terrains (e.g. Binns 1962, Davidson 1968, Green and Ringwood 1967, p. 826). The principal interest of the pyroxenes lies in their intragrain inhomogeneity. In both lithologies studied the ortho- and clinopyroxene grains commonly show (100) exsolution lamellae developed in patches (domains) or throughout the grain, although some grains show no exsolution. These features are well represented in the analysed microbe sections. Multiple spot analyses have been used to determine the compositions of single-phase material, and area scans to estimate the bulk compositions of two-phase iniergrowths. The lamellae are too fine, and usually too closely spaced, to permit separate analysis of lamellae and host, though a few spot analyses of clinopyroxene host material have been obtained. The most obvious and significant aspect of the pyroxene variation is the Ca content, reflecting widening of the solvus with cooling. Exsolved grains and domains show the highest-T compositions (i.e., highest Ca in orthopyroxenes and lowest Ca in clinopyroxenes; see analyses labelled A in Table 2). The lowest-T compositions are found in optically homogeneous grains and domains (i.e., lowest Ca in orlhopyroxenes and highest Ca in clinopyroxenes; see analyses labelled B in Table 2). However, there is apparently no hiatus between the higher- and lower-T compositions because intermediate compositions are found in exsolved and optically homogeneous material. In some instances the lower-T compositions are confined to grain margins, but usually there is no obvious regular zonal pattern in the variations. In view of the compositional relations outlined above, the exsolved grains and domains are interpreted as relics of earlier-generation, higher-T, single phase pyroxenes that exsolved during cooling. The homogeneous, single- phase grains and domains arc regarded as the product of textural and chemical readjustment of these earlier, highcr-T pyroxenes. The present variation in the chemical compositions of the pyroxenes, within and between grains, shows that final adjustment (equilibration?) was grossly incomplete, being achieved in <50% of the pyroxene analysed. Table 2 Microprobe analyses and cation proportions for Pyroxenes. Hornblendes and Biotites Sample No. Orthopyroxenes Clinopyroxenes Hornblendes Biotites 103a 282 103a 282 103a 282 103a 282 At Bt A B A B A B C E Si02 47.87 48.00 48.11 48.28 49.48 49.82 49.81 50.25 40.67 41.46 34.71 34.63 35.50 Ti02 0.22 0.16 0.23 0.19 — 0.14 0.10 0.04 1.73 1.86 4.95 4.82 4.96 AI 2 O 3 0.84 0.81 0.89 0.81 1.42 1.25 1.23 1.19 11.04 10.39 13.74 14.56 13.44 Fe 203 * 0.88 0.89 0.93 0.89 1.87 1.28 1.72 1.15 3.40 2.56 FeO* 38.90 39.12 37.49 37.93 19.74 18.94 19.12 18.11 22.19 21.06 27.62# 26.17# 25.47# MnO 1.35 1.39 1.16 1.16 0.70 0.60 0.57 0.49 0.21 0.18 — — — MgO 8.63 9.02 10.06 10.02 7.54 7.17 8.04 7.96 5.24 6.62 5.97 6.78 7.75 CaO 1.43 0.75 1.23 0.83 19.13 20.85 19.43 20.82 11.04 11.14 — — — Na20 — — — — 0.17 0.15 0.16 0.15 1.27 1.40 — — — K 2 O — — — — — — — — 1.59 1.65 9.53 9.58 9.41 Total 100.12 100.14 100.10 100.11 100.05 100.21 100.18 100.16 98.37 98.32 96.53 96.55 96.53 Cations 6 oxygens 23 oxygens 22 oxygens Si 1.959 1.962 1.952 1.960 1.946 1.953 1.949 1.960 6.324 6.402 5.484 5.428 5.536 Aliv 0.041 0.038 0.043 0.039 0.054 0.047 0.051 0.040 1.676 1.598 2.516 2.572 2.464 Al'^' — 0.001 — — 0.012 0.011 0.006 0.015 0.348 0.293 0.044 0.118 0.006 Ti 0.007 0.005 0.007 0.006 0.004 0.003 0.001 0.202 0.216 0.589 0.568 0.582 Fe^+ 0.027 0.027 0.028 0.027 0.055 0.038 0.051 0.034 0.397 0.297 Fe2+ 1.331 1.337 1.272 1.288 0.649 0.621 0.626 0.591 2.885 2.720 3.649 3.430 3.322 Mn 0.047 0.048 0.040 0.040 0.023 0.020 0.019 0.016 0.028 0.024 Mg 0.526 0.550 0.609 0.606 0.442 0.419 0.469 0.463 1.214 1.523 1.406 1.584 1.801 Ca 0.063 0.033 0.053 0.036 0.806 0.876 0.815 0.870 1.839 1.843 — — — Na — — — 0.013 0.011 0.012 0.011 0.383 0.419 K — — — — — — — — 0.315 0.325 1.922 1.916 1.872 Zz 2.000 2.000 1.995 1.999 2.000 2.000 2.000 2.000 8.000 8.000 8.000 8.000 8.000 5Ix+y 2.001 2.001 2.009 2.003 2.000 2.000 2.001 2.001 7.611 7.660 7.610 7.616 7.583 Mg/Mg +Fe2+ 0.283 0.291 0.324 0.320 0.405 0.403 0.428 0.439 0.296 0.359 0.278# 0.316# 0.352# Ca':.....: 3.3 1.7 2.8 1.9 42.5 45.7 42.7 45.2 Mg: 27.4 28.6 31.5 31.4 23.3 21.9 24.6 24.1 Fe2+ 69.3 69.7 65.8 66.7 34.2 32.4 32.8 30.7 * Fe 203 and FeO calculated by the method of Papike (1974). U Total Fe as FeO. t A- high-T compositions; B - low-T compositions (see text). 98 Journal of ihc Royal Society of Western Australia. Vol. 68. Part 4. 1986 The lowest-T compositions of coexisting pyroxene pairs (i.e., those showing the widest solvus gap) have been applied to the two-pyroxene thermometer to estimate the temperature of final adjustment. Inherent in this approach is the perhaps questionable assumption that the observed lowesi-Ca orthopyroxene and hi^est- Ca clinopyroxene achieved mutual equilibrium. Although the highest-T bulk compositions of exsolved domains cannot be confidently regarded as precisely equivalent to those of the high-T single-phase precursors (because of the possibility of significant intragranular migration of the exsolved phase), these bulk compositions have been used to roughly estimate the temperature of the high-T ‘event’. The results obtained from several versions of the two-pyroxene thermometer are shown in Table 3. Although the two samples yield mutually consistent temperatures for each particular version of the thermometer, poor agreement between the various versions renders the results uncertain. The Wells (1977) and Wood and Banno (1973) versions are generally believed to overestimate metamorphic temperatures by 50-100° C, so the most probable equilibration (?) temperatures for these Albany pyroxenes arc 750-800“ C for the peak T and 600-650° C for the final adjustment. Final adjustment occurred after 50-160° C of cooling below the peak T, according to most versions of the thermometer (Table 3). Table 3 Temperatures {°C) derived, using several versions of the two-pyroxene thermometer, fVom pyroxene pairs in enderbitic gneiss and mafic granulite, Albany. W.A. HighT Low T Sample No. 103a 282 Mean# 103a 282 Mean# Wells (1977) 920 903 912 815 840 828 84 Wood & Banno (1973) 843 831 837 775 791 783 54 Kretz (1982) Kp ... 708 796 752 748 742 745 7 Lindsley (1983)* cpx 820 790 805 640 660 650 155 Ross & Huebner (1975) 780 780 780 600 640 620 160 Kretz (1982) Solvus 735 742 739 586 628 607 132 Lindsicy (1983)» opx 800 740 770 600 610 605 165 • P - 5 kbar # Average for 1 03a and 282 ••Difference between mean high T and mean low T Opaque oxides Ilmenite shows no visible “exsolution” lamellae in either sample. Individual grains are internally fairly uniform and variation between grains is only small. The analyses in Table 4 are average compositions derived from numerous spot analyses of 10 grains in each sample. The observed ranges in average compositions for individual grains are 1.78 — 2.47 wt % MnO and 4.0 — 5.3 mol % Hem for sample 103a. and 0.79 — 1.78 wt % MnO and 3.4 — 5.7 mol % Hem for sample 282 (where llm:Hem proportions are calculated by Carmichael’s ( 1 967) method). The significance, if any, of the small observed variations with respect to cooling history is not known. Titanomagnetiie is much less common than ilmenite. In sample 103a a few grains show ilmenite “exsolution” lamellae parallel to one set of (1 1 1) planes of the host, but most are homogeneous. The ilmenomagnetite intergrowths have bulk compositions (determined by area scans) with relatively high Usp contents up to 49.5 mol % (Table 4, analysis Al). The magnetite host contains about 3.3 mol % Usp (analysis .A2) and the ilmenite lamellae are too fine for separate analysis. In contrast, the titanomagnetite grains lacking “exsolution” lamellae have intermediate Usp contents. Individual grains are fairly homogeneous, but variation between grains is significant, namely 6.0 — I L9 mol % Usp (analyses E and C). Si02, AI2O3, V2O3 and MgO are ubiquitous minor constituents. Titanomagnetite was not encountered during probe analysis of sample 282. Application of the magnetite-ilmenite thermometer to sample 103a is complicated by the variation in titanomagnetite and ilmenite compositions, and by consequent difficulty in recognising possible equilibrated compositional pairs. The range in titanomagnetite compositions is thought to be the result of variable readjustment during cooling. The relatively small variation in recorded ilmenite compositions suggests either (i) a close approach to thorough equilibration was achieved by this mineral at some stage, or (ii) variation in T-foT conditions during cooling closely paralleled the Ilm-Hem curves on the T-fo2 diagram (e.g., Spencer and Lindsicy 1981), so that there was only a small variation in the stable composition of Ilm-Hem solid solutions during cooling. The most likely instance of chemical equilibrium between titanomagnetite and ilmenite is provided by the only pair of grains observ'ed in mutual contact. These grains are fairly homogeneous and contain 1 1.9 mol % Usp and 4.6 mol % Hem, respectively. These compositions applied to the curves defined by Spencer and Lindsicy (1981, Fig. 4) suggest T~610°C and fo2''-19.4 log units. The minimum recorded Usp and Hem contents (3.3 and 4.0 mol %, respectively) may possibly be used to give a very rough indication of the conditions at which retrograde adjustment ceased; namely -550°C and ~- 21.7 log units fo2* Because of their high bulk Usp contents, the ilmenomagnetite intergrowths are interpreted as relics of higher-T titanomagnetite formed possibly during the peak of metamorphism. Estimation of the T of formation of this hi^er-T precursor is precluded by lack of compositional data on the coexisting ilmenite. However, if it is assumed that coexisting ilmenite contained a Hem content at least equal to the presently observed maximum (i.e.. > 5.3 mol %), then a temperature > 750°C is indicated by the Spencer and Lindsley curves. Hornblende The hornblendes are magnesian hastingsitic or ferroan pargasiiic hornblendes (depending on the accuracy of the Fe^+ calculation). They have fairly high Fe/Mg ratios, reflecting the host-rock compositions, but are otherwise similar to typical granulite-facies hornblendes elsewhere (cf. Engel and Engel 1962, Binns, 1965, Davidson 1971), including the surrounding south-coast region of the Albany-Fraser Province (cf Stephenson 1977). The analyses presented in Table 2 are averages of numerous spot analyses of 5 and 14 grains in samples 103a and 282, respectively. In both samples there is relatively little 99 Journal of the Royal Society of Western Australia. Vol. 68. Part 4. 1986 Table 4 Microprobe analyses and cation proportions for opaque oxides Sample No. Ilmenites# Magnetitest 103a 282 103a Al A2 C E Si02 — — 0.40 0.08 0.34 0.12 Ti02 49.99 50.05 16.99 1.04 3.71 1.92 AI2O3 — — 0.79 0.60 1.48 1.16 V2O3 — — 0.25 0.33 0.37 0.39 Fe203* 5.09 4.97 34.20 65.74 59.02 63.29 FeO* 42.78 43.60 46.47 32.21 34.83 32.96 MnO 2.15 1.39 0.68 — — — MgO 0.24 — 0.24 0.15 Total 100.01 100.01 100.02 100.00 99.99 99.99 Cations 3 oxygens 4 oxygens Si _ 0.015 0.003 0.013 0.005 Ti 0.952 0.953 0.480 0.030 0.106 0.055 Al 0.035 0.027 0.066 0.054 V 0.007 0.010 0.011 0.012 fe^+ 0.097 0.095 0.967 1.897 1.685 1.815 Fe2+ 0.905 0.923 I.46I 1.033 1.105 1.051 Mn 0.046 0.030 0.022 — — ^ 0.014 — 0.014 0.009 / Cations 2.000 2.001 3.001 3.000 3.000 3.001 Hem mol % 4.85 4.73 Usp mo 1 % — — 49.52 3.28 11.88 5.98 * Fe203 and FeO calculated by the method of Carmichael ( 1 967). Magnetite analyses recalculated on the ulvospinel basis. # Ilmentite analyses are average compositions. t Magnetite analyses; Al - bulk composition of ilmenomagnetite with maximum recorded Usp content; A2 = magnetite host in Al; C and E - homogeneous titanomagnetite grains with maximum and minimum recorded Usp contents, respectively. Thermometer Temperature 500° 600° 700° 800° Opx - Cpx 1 1 1 1 )— ?~l Mag - llm K_?_| h*? Cpx - Hbe Hbe - Flag H— ? , Flag - K-feld \ ^ i ^ Figure 4. — Summary of temperature estimates derived from various thermometers. ‘Hie uncertainty brackets shown result from variations between different versions of the thermometer (Opx-Cpx), intrasample variations in mineral compositions (Mag-Um). differences between samples (Cpx-Hbe), and imperfect knowledge of the effects of structural states and minor components (Hbe-Plag and Plag-K-feld). variation within or betw'een grains suggesting, in contrast to the pyroxenes, a possible close approach to equilibrium. A careful search was made in sample 103a for internal zoning in hornblende overgrowths and poikiliblasis, and for variations related to the identity of the overgrown or included phase. The small variations observed showed no systematic patterns. The compositional differences in hornblende between the two samples are small; in 282 the Mg + Fe^+ ratio is higher, reflecting differences in host-rock and pyroxene compositions, and the Al content is a little lower. Kretz and Jen (1978) suggested that the distribution of Mg and Fe^^ between coexisting clinopyroxene and hornblende has potential as a thermometer. In the present samples the distribution coefficient = 1.61 for 103a and 1.40 for 282, where [X/(l- X)p*. [(I-XyXJhb, X = Mg/(Mg+Fe2r). and the clinopyroxene compositions used are the low-T ones. The difference between the values for the two samples is too large to be explained by the small difference in Al’^' in the hornblendes (see Kretz and Jen ' 1978. Fig. 1). The results suggest clinopyroxene- homblendc equilibration temperatures of -690‘*C for 1 03a and -760X for 282. according to the calibration of ' Kretz and Jen (1978, Fig. 2). These temperatures are broadly consistent with the two-pyroxene temperatures, but must be viewed with caution because (i) uncertainties in the calculated Fe-^‘ contents, especially for the hornblendes, are large enough to produce significant uncertainties in the Kq values; (ii) the clinopyroxene and hornblende compositions may not represent mutual equilibrium, in which case the derived temperatures are meaningless; and ('*•) K-DMg pe recognised as a well tested, reliable thermometer. Biotile Biotile is a very minor component of both samples, so the average compositions presented in Table 2 are based on a small number of spot analyses. The high Fe/Mg ratios are consistent with the pyroxene and hornblende values. Variation within grains is small, but in sample 1 03a there are significant differences between grains in widely separated parts of the analysed thin section (compare analyses 103a C and E, Table 2). 100 Journal of the Royal Society of Western Australia. Vol. 68. Part 4. 1986 Feldspars Plagioclase is generally optically homogeneous, though in the enderbitic gneiss some grains are antiperthic. In both the analysed samples, plagioclase compositions are uniform within and between grains. The mean compositions derived from spot analyses are ^^60.3^^37.2^2.5 103a and Na 6 i. 8 ^^ 36 .i^ 2 .i 282 (Table 5). Rims may appear to be slightly more or less calcic than cores but the difference is usually less than I mo! % An. Such small and inconsistent differences are probably largely the result of analytical imprecision, and the plagioclase is therefore regarded as unzoned. Bulk compositions of antiperthilic grains were not measured because the concentration of K-feldspar blebs is highly variable, and therefore the pre-exsolution composition of the single-phase precursor is not readily estimated. These antiperthitic domains are presumably relics of the peak metamorphic assemblage. Table 5 Microprobe analyses and cation proportions for feldspars Sample No. Plagioclases K-feldspars 103a 282 103a 282 Si02 59.30 59.76 63.22 63.63 AI 2 O 3 25.71 25.43 18.71 18.70 CaO 7.69 7.44 BaO 2.17 1.81 Na20 6.88 7.03 0.75 0.61 K 2 O 0.43 0.36 15.15 15.26 Total 100.01 100.02 100.00 100.01 Cation proportions based on 8 oxygens Si 2.647 2.663 2.962 2.971 AJ 1.353 1.336 1.033 1.029 Ca 0.368 0.355 Ba 0.040 0.033 Na 0.595 0.607 0.068 0.055 K 0.025 0.020 0.905 0.909 z 4.000 3.999 3.995 4.000 0.988 0.982 1.013 0.997 Na: 60.3 61.8 6.7 5.5 Ca; 37.2 36.1 — — Ba: — — 3.9 3.3 K 2.5 2.1 89.4 91.2 K-feldspar is essentially orthoclase in both the analysed samples, though way>' extinction in some grains in 103a suggests incipient inversion to microcline. Some grains in 103a arc weakly perthitic, but generally there is little sign of exsolution. Multiple spot analyses indicate uniform compositions around Na^ 7 Ba 3 9 Kg 9 4 in I03a and Na 55 Ba 33 K 9 i 2 in 282 (Table 5). The relatively high Ba contents are reflected in the whole- rock analyses (Table 1). Bulk compositions of perthitic grains were not determined because, as with the antiperthite, the concentration of albite blebs is highly variable. Application of the feldspar compositions to any version of the two-feldspar thermometer yields temperatures well below those indicated by the two- pyroxene thermometer, and also below the usually accepted range for the granulite facies. For example, the Stormer (1975) version yields temperatures around 400- 450“ C (assuming P = 5 kbar, Stephenson 1984); i.e. below the accurately calibrated range of the method. The Whitney and Stormer (1977) version, which attempts to take into account the low-T structural states (but see criticism by Brown & Parsons (1981)), gives slightly higher temperatures, but still < 500“ C. Another element of uncertainty is introduced by the unknown effect of Ba in K-feldspar on the distribution of Na, and hence on derived temperatures. If Ba plays the same role as Na, an even higher T might be inferred, but probably still < 550“ C Despite the uncertainty surrounding the two-feldspar thermometry it is clear that the Ab contents of the K- feldspars are much lower than usual for the granulite facies and, because of the scarcity of albite exsolution lamellae, there is no clear evidence of an origin involving granulite-facies temperatures. Therefore, either exsolved albite has migrated out of the K-feldspar during cooling (to be incorporated in coexisting plagioclase?), or the K-feldspar was metasomatically introduced at a T substantially below the metamorphic peak. Weakly developed replacement textures shown by K-feldspar in the enderbitic gneiss and mafic granulite inclusions may support the latter suggestion. However, the high Ba contents in these K-feldspars are not found in undoubted metasomatic K-feldspars in meta.somatised enderbitic gneiss near its contact with the Albany .Adamellite (cf. Stephenson 1974). For this reason the K-feldspar in samples 103a and 282 is not thought to be metasomatic in origin, and therefore the present K-feldspar (and plagioclase?) compositions are probably the result of exsolution (and re-equilibration?) during cooling. The chemical uniformity of coexisting hornblende and plagioclase could imply a close approach to equilibrium between these minerals. Their measured compositions applied to the amphibole-plagioclase thermometer developed by Spear (1980) suggest a T close to 530“ C for both samples; i.e. significantly lower than the homblende-clinopyroxene temperatures, but comparable with the roughly estimated two-feldspar temperature. However, the reliability of the amphibole- plagioclase thermometer is severely limited, especially when amphibole Fe^+contents are calculated. Spear (1980) suggested an uncertainty of ± 50“ C. Synthesis and discussion At least two factors preclude a precise reconstruction of the cooling history of the rocks in question. Firstly, many of the reactions by which retrograde adjustments were achieved cannot be identified with confidence, though some resonable inferences can be drawn. Secondly, it is not possible to determine precisely the temperatures at which the various readjustments occured, partly because of uncertainties in the relevant thermometers, and partly because of the difficulty in recognising mutually equilibrated compositions in partially or completely readjusted associated minerals. Indeed, it is possible that none of the present mineral compositions represents chemical equilibrium with respect to the ambient conditions at any particular stage, although this is perhaps an undulv pessimistic proposition in view of the probable slowness of post- metamorphic cooling. Anyway, the reconstruction outlined below and summarised in Fig. 4 must be regarded as speculative and, at best, only semi- quantitative. 101 Journal of the Royal Society of Western Australia. Vol. 68. Part 4. 1986 The microprobe data indicate that orthopyroxene, clinopyroxene and titanomagnetite certainly participated in relatively high-T chemical readjustment during cooling, and that feldspars very probably did. Participation of hornblende, ilmenite and biotite remains unproved, but likely. Variation in pyroxene compositions are interpreted as the result of partial chemical readjustment, essentially involving Ca-Fe^^-Mg redistribution defined by the pyroxene solvus, dunng the early stages of cooling. Although agreement between the different versions of the two-pyroxene thermometer is poor, it appears that this readjustment occurred over a cooling interval of about 50- 1 SO” C. The peak T recorded by the pyroxenes was around 750-800" C. Readjustment may have continued to a temperature as low as 600-650° C, according to several versions of the thermometer. Similar results have been recorded in pyroxene pairs from Cape Riche, about 90 km northeast of Albany, except that re-equilibration was more thorough and apparently occurred at a higher T (700-800° C) at Cape Riche (Stephenson 1984). The magnetite-ilmenite thermometer yields temperatures broadly consistent with the two-pyroxene results. The uncenainties are large (see above), but the titanomagnetite appears to record a peak T > 750" C, followed by readjustment at about 600" C and in some grains possibly as low as 550“ C. This suggests that the pyroxenes and opaque oxides readjusted over roughly the same T range, although opaque oxides possibly continued to respond to cooling for a short interval after the pyroxenes had “closed’'. The textural occurrence of hornblende in the enderbitic gneiss and many of the mafic granulite inclusions (though not 282) strongly suggests retrograde growth (see above). This hornblende shows brown-green absorption colours, medium gram size and smooth grain boundaries against pyroxenes. These arc features typical of granuliie-facies hornblendes, contrasting with the blue-green absorption colours, acicular-aggrcgaie form and ragged grain boundaries typical of lower-T retrograde calcic amphiboles. The relatively high, contents of A-sile (‘edenite’) alkalis and low Ah' are also characteristic of granulite-facies hornblendes (cf. Binns 1965, Fig. 6). Thus, hornblende growth is believed to have occurred during the early stages of cooling, still within the T regime of the lower granulite facies, probably according to a reaction such as Hbe ^ Qtz Opx ^ Cpx Flag * H 2 O ( 1 ). Experimental results of Binns ( 1 969) and Spear (1981) show that the temperature interval encompassed by this mullivariant equilibrium substantially overlaps the T interval inferred above for pyroxene readjustments, especially for Si02-ovcrsaturaled rocks and Ph 20 (factors which extend the stability of pyroxenes to lower temperatures (e.g. Spear 1981)). Therefore pyroxene and opaque oxide readjustments and hornblende development were probably inter-related, synchronous processes occurring within the T regime of the lower granulite facies. The T estimates derived from the hornblendc-clinopyroxene thermometer tend to support this interpretation. The relatively low T 530°C) indicated by the amphibole-plagioclasc thermometer is well below the minimum T at which reaction (1) is likely to equilibrate (see Spear 1971), and also below the T range usually associated with brown-green hornblende. Though the uncertainties inherent in the amphibole-plagioclase thermometer are severe, this suggests that final adjustment of plagioclase may have been independent of reaction (1). Support for this suggestion is provided by the low temperatures indicated by the two-feldspar thermometer; i.e. roughly 500-550“C (or less?). If the feldspars did indeed continue to adjust at temperatures below that of final adjustment of the mafic phases, then the hornblende-plagioclase temperature is meaningless. The uncertainties involved in this study serve to highlight the limitations inherent in the application of clement-distribution thermometers to reconstruct metamorphic cooling histories in rocks that preserve evidence of retrograde readjustments. These limitations stem from: (i) fundamental shortcomings in the thermometers themselves; (ii) possible failure of minerals to achieve or preserve chemical equilibrium at any particular stage (or stages) during readjustment; (iii) difficulties in recognising any mutually equilibrated adjusted compositions that may be present in coexisting minerals. Acknowledgements . — The author is indebted to A. J. Stolz and M. D. Speak for assistance with the microprobe analyses. G. I. Z. Kalocsai for some of the whole-rock analyses, J. S. Cook for polished thin section preparation, D. Windle for drafting the diagrams, and R. K. Vivian for typing the manuscript. References Beach. A. (1974). — Amphibolilization of Scourian granulites. Scott. J. Geol.. 10: 35-43. Binns, R. A. (1962).— Metamorphic pyroxenes from the Broken Hill district. New South Wales. Mineral. Mag.. 33; 320-338. Binns. R..A. (1965). — The mineralogy of metamorphosed basic rocks from the Willyama Complex. Broken Hill district. New South Wales. Part I Hornblendes. Mineral. Mag.. 35: 306-326. Binns, R..\. (1969). — Hydrothermal investigations of the amphibolite- granulitc facies boundary. Oeol. Soc lt«/. Spec hihLs.. 2 341-344. Brown. L. and Parsons. 1. (1 981 ).— Towards a more practical two- feldspar geothermomelcr. Conir Mineral. eelrolog}\ 76; 369-377. Carmichael. I. S. H. (1967),— The iron-t»ianium oxides of salic volcanic rocks and their associated ferromagnesian silicates. Contr. Mineral. Petrology. 14: 36-64. DavidMtn, L. R. (1968). — Variation in ferrous iron-magnesium distribution coefftcients of metamorphic pyroxenes from Quairading. Western Australia. Contr. Mineral. Petrology. 19; 239-259. Davidson. L. R. (1971). — Metamorphic hornblendes from basic granulites of the Quairading district. Weslcm Australia. Seucs Jarh Mineral. Monaish., 344-359. Pngcl. A. E. J. and Engel C. G. (1962). — Hornblendes formed during progressive mciamorphism of amphibolites. northwest Adirondack Mountains. New York. Geoi Soc. Amer. Bull. 73: 1499-1514. Green, D. H. and Ringwood. .A. E. (1967). — An e.xperimenial investigation of the gabbro to cclogiie transformation and its petrological applications. Geochtm. Cosmochim. . icla. 31; 767-833. Kretz. R. (1982). — Transfer and exchange equilibria in a portion of the pyroxene ouadrilatcral as deduced from natural and experimental data. Geochim. Cosmochim. 'iaa. 46: 41 1-421. Kretz, R. and Jen. L. S. (1978). — EiTecl of temperature on the distribution of Mg and Fc- ' between calcic pyro.xeiic and hornblende. Can. .Mineral.. 16: 533-537. Lindslev. D. H. (1983). — Pvroxene ihermometrv. .Uncr. .Mineral. 68: 477- ■ 493. Paptke. J. J.. Cameron. K. and Baldwin. K. (1974). — Amphiboles and pyroxenes: characterization of other than quadrilateral components and estimates of ferric iron from microprobe data. Geof. Soc .4mer. .Ah.sir. Program. 6: 1053-1054 Ross, M. and Huebner. J. S. (1975). — .A pyroxene geothermometer based on composition-temperature relationships of naturally occurring onhopyroxene, pigeonile. and augite. E.xiended Abstr. International Conference on Geothermomeirv and Barometry. Pennsylvania Slate Univ. Sills, J. D. (1983). — Mineralogical changes occurring during the retrogression of Archaean gneisses from the Lewisian complex of NW’ Scotland. Lithos. 16; 1 13-124. 102 Journal of the Royal Society of Western Australia. Vol. 68, Part 4, 1986 Spear. F. S. (1980). — NaSi 5=^ CaAl exchange equilibrium between plagioclasc and amphibole. Conlr. Mineral. Petrology. 72: 33-4 1 . Spear, F. S. (1981). — An experimental study of hornblende stability and compositional variability in amphibolite. Amer. J. Sci.. 281: 697- 734. Spencer. K. J. and Undsley. D. H. (1981). — A solution model for coexisting iron-titanium oxides. Amer Kfineral.. 66: 1 189-1201. Stephenson. N. C. N. (1974). — Petrology of the Albany and Torbay Adamellite plutons, near .Albany, Western Australia. J. Geo/. Soc. Ausi: 21: 219-246. Stephenson. N. C. N. (1977). — Coexisting hornblendes and biotiles from Prccambrian gneisses of the south coast of Western Australia. Lilhos. 10: 9-27. Stephenson, N. C. N. (1979). — Coexisting garnets and biotites from Prccambrian gneisses of the south coast of Western Australia. Lilhos, 12: 73-87. Stephenson, N. C. N. (1984). — Two-pyroxene thermometry of Prccambrian granulites from Cape Riche, Albany-Frascr Province. Western Australia. / Met. GeoL. 2: 297-314. Stormer. J. C. (1975). — A practical two-feldspar geothermometer. Amer. Mineral., 60; 667-674. Ware. N. G. (1981). — Computer programs and calibration with the PIBS technique for quantitative electron probe analysis using a lithium- drifted silicon detector. Computors d Geosciences, 7: 1 o7- 1 84. Wells. P. R. A. (1977). — Pyroxene thermometry in simple and complex systems. Contr. Mineral. Petrology. 62; 129-139. Whitney. J. A. and Sloimer. J. C. ( 1 977). — The distribution of NaAlSi30g between coexisting microcline and plagioclasc and its effect on gcothcrmomelric calculations, .imer. Mineral.. 62; 687-691. Wood. B. J. and Banno. S. (1973). — Garnci-onhopyroxcne and orthopyroxene-cUnopyroxenc relationships in simple and complex systems Contr. Mineral. Petrology, 42: 109-124. 103 'Mm