Igneous Rocks from Western Samoa.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 57, 1927, Page 254

Igneous Rocks from Western Samoa. By J. A. Bartrum, Auckland University College. [Read before the Third Science Congress of the New Zealand Institute, Dunedin, 28th January, 1926; received by Editor, 29th February, 1926; issued separately, 7th December, 1926.] In 1920 Dr. J. A. Thomson, Director of the Dominion Museum, collected a number of igneous rocks from various parts of Western Samoa, and very kindly forwarded about thirty-four thin slices cut from these rocks, and from others collected by Mr. F. Wood in 1915, to the writer for description. From one cause or another the work has been unavoidably delayed, but this delay has proved to be an advantage in that a very interesting and instructive account of the geology of American Samoa by Professor R. A. Daly (1924) has been published since study of the collection was begun. The writer is greatly indebted to Dr. Thomson for his courtesy in allowing him to describe this Samoan collection, and for arranging with the Dominion Laboratory for analysis of four of the rocks. Mr. F. T. Seelye, who undertook the analyses, has also very kindly furnished the norms and classification of the rocks. Thanks are due also to Dr. P. Marshall and Professors C. A. Cotton and R. Speight for the loan of Samoan literature not available in Auckland. The collection does not include any rocks that shew outstanding difference from those previously described from Samoa. On the whole there is surprising uniformity of type. The great majority of the slides represent olivine-rich basalts, but there is every gradation from picritic and possible limburgitic basalts to types with relatively little olivine but plentiful plagioclase (generally acid labradorite), which chemically approach the border-line between basalts and andesites. There is far less variety of type than has been described by Daly (1924) from American Samoa. The only rock not belonging to the basalt series is a soda-trachyte collected from beach boulders at Masina, Fangaloa Bay, Upolu Is. Since so little fresh information has been added by the present study, any very detailed description of individual rocks would be out of place, and has not been attempted. General Account of the Basalts. All of the basaltic rocks are very closely-related members of a series readily derivable by minor differentiation from a single magmatic intrusion. The more basic types include rocks closely allied to the picritic basalts described by Cross (1915) and Washington (1923) from Hawaii. Unfortunately one of the two analyses made of representatives of these picrite basalts is somewhat vitiated by the presence of a comparatively large amount of calcite filling vesicles of the rock. The majority of the rocks are olivine basalts in which plagioclase is not prominent, but, by increase in the amount of this latter mineral, these gradually pass into types in which

plagioclase is abundant. These plagioclase-rich basalts are not, however, nearly as well represented in the present collection as in that of Friedlaender described by Weber (1909). The basalts typically are strongly porphyritic; the phenocrystic minerals are characteristically olivine and pyroxene, to which must be added plagioclase in certain less basic rocks. In one specimen (No. 2) collected from a beach boulder at Samamea in Fangaloa Bay, Upolu, plagioclase is especially plentiful and is developed not only in the groundmass but also as tabular phenocrysts as much as 1 cm. in length. The pyroxene of the rocks is generally a pale greenish-grey augite or diopside, though in a very large number it is pleochroic violet titaniferous augite. In a dense non-porphyritic rock (No. 13) from Uafato Bay, Upolu, there is a very small unimportant quantity of a rhombic pyroxene determined as hypersthene, though it is in such small crystals that the determination is a little doubtful. It will be remembered that Jensen (1906) described an olivine-enstatite basalt from Savaii, but the two rocks are entirely dissimilar one from the other. There are numerous small flakes of biotite in the rock (No. 5) from Salimu in Fangaloa Bay. This mineral has been recognized in a number of other Samoan basaltic rocks by several other writers such as Moehle (1901, p. 4), Weber (1909, p. 294) and Daly (1924, p. 102). The groundmass of the basalts is very often more or less pilotaxitic, and fluxional arrangement of the minute lath-like crystals of plagioclase is very well shewn in some instances (See Fig. 1). There is, however, an expectable development of glass in several types, and one (No. 20), which obviously has cooled rapidly during solidification, has a groundmass which is completely vitrophyric. The grain-size is subject to considerable variation. There are a few coarse-grained doleritic basalts (Nos. 14, 31) and, at the other extreme, quite a number of finely microcrystalline varieties (Nos. 13, 17, 19, 21, 30) which can only be studied by the use of very high-power lenses (See Figs. 2 and 3). In mineral constitution the groundmass consists essentially of plagioclase, a pyroxene, which is either diopside or a titaniferous augite, and oxide iron ores. Of these latter magnetite is more prevalent than ilmenite. The average content of these iron ores is about 9 or 10 per cent., though occasionally it is considerably in excess of this. The proportion of recognizable apatite is variable. It is usually inconspicuous, but in places occurs in very numerous minute needle-like prisms. It is often very difficult to obtain the variety of plagioclase in the groundmass, especially where fluxionally arranged laths are cut in unfavourable directions. In many slides, in addition to being very small, these “laths” are irregular in outline and yield poorly-defined albite twin lamellae. Where accurately determinable the variety is generally andesine-labradorite or acid labradorite, and is seldom more calcic than Ab45 An55. In two instances (Nos. 25 and 34) the groundmass is unusual in that amongst the last-crystallized material there are much larger very poorly-defined crystals of a mineral which envelops most other constituents, including the usual small lath-like crystals of

labradorite. It was at first suspected that nepheline was present, though in some instances cleavage is visible and with this there is obtainable an extinction angle which, with the refractive index, indicates that the material is in part a plagioclase decidedly more acid than the labradorite of the laths it enwraps. Confirmation of the suspicion that nepheline is also present was sought in staining tests, but these proved unsatisfactory, and recourse was had to digests of the finely powdered rock wih hydrochloric acid and subsequent evaporation of the solution. No cubes of sodium-chloride were obtained, so that nepheline apparently is not present, but it is by no means certain that plagioclase is the only mineral represented in this enveloping medium. Detailed Description of Basalts. 1. Picrite Basalts and ? Limburgites. A striking example (No. 4) of picrite basalt was collected from a beach boulder at Masina, Fangaloa Bay, Upolu. In hand-specimen it appears to be closely crowded with phenocrysts of olivine and augite up to about 1 cm. in diameter, and the microscope section (See Fig. 4) indicates that the phenocrystic material is only slightly less in amount than the groundmass, for a micrometric analysis gave the volume percentages as follows:—olivine (phenocrysts) 20.2; augite (phenocrysts) 20; calcite (in vesicles) 2.7; groundmass 57.1. As the analysis (No. 2, Table 1) shews, the percentage of calcite in the material analysed is much higher than that estimated in the vesicles, and, as the rock is unweathered and the calcite seems to be entirely restricted to the vesicles, it appears that the fragments selected for analysis were especially rich in this mineral. The olivine is fairly sharply idiomorphic and shews only incipient alteration to serpentine. The augite is a zonally-built pinkish-brown titaniferous variety present in large moderately euhedral crystals. It has numerous minute inclusions of plagioclase, magnetite, occasional calcite and olivine and rare brown hornblende as well as some larger ones consisting mainly of pale-brown glass. The rock shews also a few microphenocrysts of moderately basic labradorite as much as 1.5 mm. in length. A Rosiwal determination of the proportions of the minerals of the finely crystalline groundmass was somewhat unsatisfactory, but indicated a greater proportion of pyroxene relative to plagioclase than is normatively present. The pyroxene includes both titaniferous augite and diopside. Magnetite constitutes about 17 per cent. of the groundmass, and needles of apatite are plentiful. The normative plagioclase is andesine-labradorite; modally the variety appears to be a little more calcic though precise determination is difficult. The name picrite basalt has been used for this and similar rocks of this group, following its use by Cross (1915, p. 29.) for a similar rock from Haleakala, Hawaii. The chemical analyses of the two rocks are compared in Table 1, that for the Samoan rock being re-calculated on account of the large amount of calcite noted in both mode and norm. In describing several “chrysophyric” basalts and picrite basalts from Hawaiian localities, Washington (1923) adopts as a means of separating the picrite basalts from other

Fig. 1. Olivine basalt (No. 3) with fluxionally arranged laths of plagioclase in groundmass. A phenocryst of olivine is visible, but none of plagioclase though many are present in the slide. Magnification 50 diams. 2. “Doleritic” olivine basalt (No. 14). The abundance of titaniferous augite is well exemplified. Magnification 35 diams. 3. Very fine-grained non-porphyritic basalt (No. 13) with microphenocrysts of plagioclase. Magnification 170 diams. 4. Picrite basalt (No. 4) shewing phenocrysts of olivine and augite and part of an amygdule of calcite. Magnification 35 diams.

Fig. 5. Picrite basalt close to olivine basalt (No. 8) with stellate groups of titaniferous augite and phenocrysts of olivine (white) plainly visible. Magnification 35 diams 6. Phenocrysts of plagioclase and augite in an olivine basalt (No. 2) rich in large phenocrysts of plagioclase. Nicols crossed; magnification 35 diams. 7. Anorthoclase in anorthoclase trachyte (No. 1). A corrosion border is evident. Nicols crossed; magnification 35 diams. 8. Probable albite with outgrowth of more calcic plagioclase from anorthoclase trachyte (No. 1). Trachytic structure of the groundmass can be recognized. Nicols crossed; magnification 50 diams.

olivine-rich basalts the criterion that in the former (picrite basalts) the femic minerals “dominate” over the feldspars. That is the lower limit of the percentage of femic minerals is fixed at 62.5. Apart from satisfying this criterion, the Samoan picrite basalts here described as such are not very similar to those described by Washington from Hawaii, for they have a greater proportion of phenocrystic augite and less olivine. As pointed out clearly by Washington (1923, p. 470), however, the work of Bowen and Anderson shews that the mutual proportions of these two minerals depend very largely upon the rate at which crystallization has proceeded. Another rather different picrite basalt (No. 22) comes from an outcrop at the shore-line at Suifaga, Savaii. An approximate micrometric determination of the mode gave the percentages of the constituents by weight as olivine 20, augite 59.5, labradorite (Ab45 An55 approx.) 11.0, magnetite and ilmenite 9.5. This is in distinct disagreement with the norm which is shewn in column 1 of Table 2, but it is probable that the discrepancy can largely be explained by the presence in modal pyroxene of lime and alumina normatively referred to anorthite. Olivine in numerous rather small idiomorphic phenocrysts generally under 1 mm. in greatest dimension, is the only phenocrystic mineral. The groundmass is holocrystalline and consists mainly of pinkish-violet titaniferous augite in small grains and prisms, along with magnetite and ilmenite and relatively long (0.15 mm.) laths of plagioclase. No nepheline was detected, though a little is shown by the norm. There are one or two other picritic basalts which may, however, be allied to the limburgites. One of them is represented by slide No. 33 cut from a specimen of vesicular pahoehoe lava collected by Mr. F. Wood. Titaniferous augite in small but stout prisms, which are often arranged in stellate manner, is the predominant constituent; it encloses small sub-euhedral phenocrysts of olivine and numerous little crystals of magnetite, and upon it is moulded a small quantity (about 15 per cent. by volume) of plagioclase in highly irregular crystals up to 1.2 mm. in breadth, which often have their margins crowded with rods of ilmenite. Particularly in certain parts of the slide, there is a small quantity of almost colourless isotropic residuum which is apparently a glass, but may be analcite. This is closely permeated by minute somewhat dendritic though rod-like growths of ilmenite passing thence into the marginal portions of the crystals of plagioclase. In slide No. 26, cut from another rock collected by Mr. Wood, there is a far greater proportion (approximately 25 per cent. by volume) of brown “glass,” and similarly there is more phenocrystic olivine than in the last rock. There is, however, no visible feldspar. It is regrettable that insufficient material was collected to enable an analysis of this rock to be made. In view of the uncertainty as to the true nature of the “glass” in these last two rocks, it is evident that classification of them as limburgitic basalts can only be tentative. Of the remaining olivine-rich basalts there are several almost on the border-line between the olivine basalts and picrite basalts. This applies especially to the rock of a dyke at Samamea, Fangaloa

Bay, Upolu (No. 8), which is somewhat similar to the types which have just been described as possible limburgitic basalts, but is holocrystalline. The augite of this rock is mainly a violet titaniferous variety built often in most characteristic groups of stellately-arranged small prismatic phenocrysts (See Fig. 5). Olivine is very plentiful in fresh idiomorphic phenocrysts up to 1.5 mm. in length. 2. Olivine Basalts. It will readily be realized after what has been stated in the last section, that several of the rocks here classed as olivine basalts contain so great a proportion of femic minerals relative to plagioclase that they closely approach the picrite basalts. Their mineralogical composition varies only slightly in different specimens. The pyroxene of many is a titaniferous augite, in others a pale greenish-grey diopsidic variety, whilst this mineral may or may not form phenocrysts. Olivine is abundant in all rocks included in this group. A few of the olivine basalts, for example No. 31 of Wood's collection, are very coarse-grained and have their augite more or less ophitically related to the plagioclase. Examples of the more basic olivine basalts are furnished by the following slides:— No. 9: a boulder from the beach at Samamea, Fangaloa Bay, Upolu. No. 17: from a well-shaft near Ofalau Hill, Mulifanui. Nos. 24, 30, 31 of Wood's collection. It should be mentioned, however, that slide No. 30 does not lend itself to very precise classification because it is a quickly-cooled type possessing a remarkably fine-grained groundmass which is almost opaque from closely-crowded iron ore. More normal olivine basalts are illustrated by the following slides:— No. 10: at west end of Samamea, Fangaloa Bay, Upolu. No. 11: beach boulder, Uafato Bay, Upolu. No. 14: west side of lagoon, Safatu Harbour, Upolu. No. 15: from the beach, south side of Manono Is. No. 16: near summit, Manono Is. No. 18: stream boulder, Leafien Creek, near Falemanga. No. 19: stream boulder, Lalia Creek, near Falemanga. No. 20: one mile from sea, Matavanu lava field, Savaii. No. 21: crater wall, Matavanu, Savaii. Nos. 23, 25, 29, 32, 34, 35, 36 of Wood's collection. Most of the above rocks differ from the more basic olivine-rich basalts previously listed only in an increased proportion of feldspar relative to femic minerals. Olivine is variable in amount, though always present more or less conspicuously. Usually it is very strikingly fresh, but occasionally exhibits reddish-brown oxidation-stains definitely oriented in parallel rod-like fashion as already noted by Moehle (1901, p. 34). Although in the slide of Matavanu lava (No. 20) a deep-brown almost opaque glass constitutes about 75 per cent. of the rock by volume, most of the rocks are more or less finely holocrystalline. Many have small vesicles which remain unfilled except in more weathered rocks, although in a few rare instances zeolites furnish a partial filling. Though the

texture of the groundmass is characteristically fine, a few rocks are coarse-grained and shew structural approach to the dolerites in their distinctly ophitic character (e.g. No. 14; see Fig. 2). In addition to the usual laths of acid labradorite with accompanying augite, iron-ore and occasional apatite, there is in slide No. 34 the enveloping material partly referable to a more acid plagioclase, and perhaps also to orthoclase or other mineral, which has already been discussed above. 3. Olivine Basalts with Plagioclase Phenocrysts.—A very interesting basalt with numerous large rather tabular phenocrysts of plagioclase (Ab30 An70) as much as 9 mm. in length in the slide (No. 2) examined (See Fig. 6), was collected from a beach boulder at Samamea, Fangaloa Bay, by Dr. Thomson. Olivine is represented by a few large pseudomorphs of deep-green serpentine and there are quite a number of rather ill-shaped phenocrysts of titaniferous augite with numerous small inclusions of magnetite, plagioclase and serpentized olivine. The groundmass is decidedly in excess of the phenocrysts and contains very abundant magnetite (about 15 per cent. by weight) along with the usual diopside and plagioclase; the latter is slightly in excess of the diopside and is apparently andesine-labradorite. Another rock from the east end of Samamea (No. 7) was amongst those forwarded for analysis, and the result is given in column 5 of Table 1, whilst in column 6, for comparison, there is the analysis of a similar Hawaiian rock described by Washington (1923, p. 493). In section this basalt is seen to possess a moderate number of large phenocrysts of augite and less numerous ones of plagioclase (acid labradorite), whilst olivine is comparatively plentiful in crystals up to 2.5 mm. in breadth; magnetite builds occasional large crystals in addition to occurring in moderate quantity in the groundmass. In this latter, plagioclase (basic andesine) is very abundant and forms about 60 per cent. of its bulk. The normative plagioclase determined by the analysis can be seen from inspection of the norm set forth in column 5 of Table 2 to be acid andesine, though modally it is more basic. Thus a more precise classification of this rock is olivine-andesine basalt. There is a moderate quantity of orthoclase shewn by the norm and this very probably occurs modally as a feldspar-like material present in small amount in the groundmass, and enwrapping other minerals there precisely as has been described for slide No. 34. The only two other rocks of the collection, which, whilst possessing a moderate quantity of olivine, have also porphyritic plagioclase, are one (No. 12) collected from the boulder beach at Uafato Bay, Upolu, and another (No. 3) from a similar beach at Masina, Fangaloa Bay, Upolu. The groundmass of both approaches in structure that termed pilotaxitic, and the abundant laths of plagioclase shew perfect fluxional arrangement. 4. Plagioclase Basalt.—A type of basalt from a beach boulder at Salimu, Fangaloa Bay (No. 6), is distinct from others of the collection in that olivine is absent and the only phenocrysts are tabular crystals of labradorite up to 2.5 mm. in length. The rock has close-spaced vesicles of the pahoehoe type and in section is poorly translucent by reason of the abundance of irregular

growths of ilmenite. These are intergrown and intermingled with the plagioclase, titaniferous augite, and magnetite of the groundmass, the plagioclase of this second generation being approximately equal in volume to the femic minerals. 5. Non-Porphyritic (Aphyric)*See Washington, H. S., 1923, p. 474. Basalts.—Only two of all the rocks of the collection fail to show clear evidence of two periods of crystallization. The first (No. 13), a beach boulder from Uafato Bay, Upolu, is a remarkably dense and fine-grained rock poorly translucent in thin section by reason of the close packing of small crystals of magnetite. There are a few microphenocrysts of this last mineral and of more or less completely serpentinized olivine. Numbers of scattered small laths of plagioclase seldom over 0.15 mm. in length perhaps also deserve to rank as microphenocrysts, for they contrast in size and clear-cut definition with the imperfectly-defined feldspathic material enwrapping the abundant minute crystals of diopside and magnetite of the general matrix (See Fig. 3). An unusual additional mineral present in infrequent small prisms 0.05 mm. up to 0.2 mm. in length is identified as hypersthene. It shews the usual straight extinction, characteristic cross-fracture and negative optical character, though no pleochroism was detected. Very rare tiny flakes of biotite were also noted. The other aphyric rock (No. 5) is unique in the collection, and comes from a beach boulder at Salimu, Fangaloa Bay. Like the last it is very dense and fine-grained and lacks true phenocrysts, although it has numerous small microphenocrysts represented by plentiful flakes of deep-brown biotite, seldom more than 0.15 mm. in length, occasional crystals of magnetite and a number of laths of plagioclase distinctly larger than those of the general matrix. Chlorite has in places replaced the biotite and provides the filling of occasional minute vesicles. It is difficult to determine the variety of plagoiclase present, but it is not less calcic than acid labradorite; an approximate micrometric analysis gave the percentage of feldspar by weight as 60, so that this basalt comes very near the border-line between andesites and basalts following the usage of Iddings and Washington,†See, for example, Washington, H. S., 1923, pp. 468–9. if it does not actually fall within the group of andesites, since the normative plagioclase is likely to outweigh the modal. If, however, the silica content be calculated from the mode, the rock falls within the usually accepted chemical limits of the basalt group. Anorthoclase Trachyte. A very interesting trachyte was found by Dr. Thomson amongst the boulders of the beach at Masina, Fangaloa Bay, Upolu. Macroscopically it is a typical porphyritic light-grey trachyte, with plentiful large phenocrysts of anorthoclase or other feldspar which are as much as 1 cm. in greatest dimension, but seldom have the euhedral tabular form so common in similar rocks. Besides the feldspar there are scattered crystals of dark ferromagnesian mineral which, though usually small, occasionally are as much as 4 mm. in breadth.

In section (No. 1) there is a trachytic groundmass, in which clear-cut laths of feldspar dominate over somewhat granular aegirine-augite and accompanying magnetite, and shew perfect fluxional parallelism; this builds about 60 per cent. by volume of the rock and in it are set a considerable number of large phenocrysts of feldspar with far less numerous small ones of more or less completely resorbed brown hornblende, of yellowish olivine and occasional magnetite. The phenocrysts of feldspar are dominantly anorthoclase; usually they have ill-defined borders and shew a narrow sub-marginal zone of corrosion beyond which there is often outgrowth of similar material (See Fig. 8). They enclose numerous irregular inclusions of deeply pleochroic aegirine-augite, magnetite, and perhaps colourless glass. Under crossed nicols they generally show the characteristic minute discontinuous and indefinite intersecting twin lamellae (Fig. 7), but in some instances they exhibit more regular albite lamellae and can then only be separated with certainty from albite by their negative optical character and small axial angle. In addition to anorthoclase there are a few large idiomorphic crystals of a plagioclase which gives the extinction angles of acid andesine, though one small fragment enclosed in a larger crystal is as calcic as andesine-labradorite. Rare crystals of albite may be present, though the observations are not as conclusive as desirable. A crystal of such probable albite is illustrated by Fig. 8; beyond the prominent corrosion-zone there is an outgrowth of a feldspar with decidedly higher indices of refraction than those of the inner mineral, and it appears therefore to be more calcic plagioclase than albite. The plane of the section in this case is unfortunately not truly normal to the composition-planes of the albite twin lamellae, but, admitting this, the extinction-angles obtained with alternate sets of such lamellae (4° and 17°) seem inconsistent with those expectable with anorthoclase, and for this reason it is suggested that the mineral is albite.*The optical character could not be ascertained. The hornblende noted already as one of the minerals of the first generation, shews deep-brown tints and strong pleochroism. It is generally present in association with abundant corrosion-products represented by grains of pyroxene intermixed with crystals of magnetite and feldspar, whilst corrosion-pseudomorphs may alone represent the mineral. Olivine is not plentiful, whilst apatite is a conspicuous accessory mineral, for it is in comparatively large bluish-grey faintly pleochroic prisms as much as 0.15 mm. in thickness. The trachytic groundmass is practically wholly of alkali feldspar in laths averaging about 0.25 mm. in length, with 8 per cent. or 9 per cent. of prisms and grains of aegirine-augite and crystals of magnetite. The majority of the laths exhibit simple twinning and have straight extinction; in all probability these are soda orthoclase. Others shew albite twinning of an irregular kind and have small extinction angles; the analysis (No. 7, Table 1) makes it almost certain that these represent anorthoclase, though the writer was unable to decide this by determination of the optical character and optic axial angle.

An interesting feature in connection with the analysis of this trachyte is that the modal olivine is reflected in the norm (See column 7, Table 2). There is great similarity between this Samoan trachyte and anorthoclase trachytes described by Skeats and Summers (1912) from the Macedon district of Victoria, and an analysis of one of these latter rocks is given in column 8 of Table 1 for purposes of comparison. Discussion of Analyses. Analyses by Mr. F. T. Seelye of the Dominion Laboratory of four of the West Samoan rocks are set forth in Table 1, with accompanying analyses of a few rocks from elsewhere which are comparable with certain of the Samoan ones. Table 2 furnishes corresponding norms and quantitative classifications. Table 1. Analyses. 1 2 3 4 5 6 7 8 Si O2 43·82 42 19 44·83 42 99 48 35 48 42 60 48 60·10 Al2O3 11·26 9·03 9 60 10·21 12·39 13·97 18 16 18 38 Fe2O3 2·39 3·03 3 22 3·01 3 27 4 17 1·99 2·22 Fe O 9·53 9·27 9·85 10·28 8·35 9·57 2·92 3·34 Mg O 13·59 13·89 14·76 14 61 8 76 4 61 0·73 1·30 Ca O 10·22 11·95 9·20 12·54 8·54 8 86 2·13 2·28 K2O 1·43 0·47 0·50 0·52 1·69 1·29 4·12 4 57 Na2O 2 08 1·77 1·88 1·40 3·28 3·30 7 27 5·30 H2O+ 0·75 1·03 1 10 1·10 0 37 0·84 0 68 0 43 H2O- 0·60 1 02 1 09 0·82 0·54 0 42 0 30 0·96 C O2 trace 2·61 nil 0·02 nil trace Ti O2 3·09 2·72 2·89 2 +2 3 48 3·25 0 39 0 69 Zr O2 nil nil nil nil nil 0 08 P2O5 0·69 0·71 0 75 0 29 0 89 0 91 0 38 0·33 S 0 02 trace trace 0 02 trace Cr2O3 0 10 0·12 0 13 0 06 0·06 nil Ni O 0·05 0 04 0·04 0·06 0·03 nil Mn O 0·25 0·19 0 20 0 17 0·22 0·17 0 17 nil Sr O 0·02 0 02 0·02 nil 0·02 0 02 Ba O 0 05 0·04 0·04 nil 0 04 0 07 Ce2O3 &c 0·03 S O3*Water-soluble salts. 0 17 trace Na Cl* 0 09 trace trace 0 09 0·04 Cl nil 0·03 nil 100·20 100·10 100·10 100 58 100·41 99·78 99·99 99 90 1. Picrite basalt (No. 22), in situ at beach, Suifaga Savaii. Seelye analyst. 2. Picrite basalt (No. 4), beach boulder, Masina, Fangaloa Bay, Upolu. Seelye analyst. 3. Recalculation of analysis 2 on the basis of absence of calcite. 4. Picritic basalt, Haleakala, Hawaii. Steiger analyst. Cross, 1915, p. 29. 5. Olivine basalt with plagioclase phenocrysts (No. 7), east end of Samamea, Fangaloa Bay, Upolu. Seelye analyst.

6. Aphyric andesine basalt, Papalele Gulch, Mauna Kea, Hawaii. Washington analyst. Washington, 1923, p. 493. 7. Anorthoclase trachyte, from beach boulder, Masina, Fangaloa Bay, Upolu. Seelye analyst. 8. Anorthoclase trachyte, eastern slope of Mount Eliza, Macedon district, Victoria. Bayley, Hall, Lewis and Topp analysts. Skeats and Summers, 1912, p. 25. Table 2. 1 2 3 4 5 6 7 8 Quartz 0 84 4·32 Orthoclase 8 34 2 78 2 95 2·78 10·01 7·78 24·46 27·24 Albite 11·00 15·20 16 15 6 29 27·77 27·77 57·11 44·54 Anorhite 17·24 15·01 15·95 20·02 14·18 19·46 4·45 9·73 Nepheline 3 69 3·12 2·56 Corundum 1·12 Diopside 22·89 18 65 19·82 32 28 18 22 15 49 2·81 Hypersthene 11 01 11·80 3·92 12·88 6·37 Olivine 24·31 18 07 19 20 24 23 11·77 2·65 Magnetite 3·48 4·41 3·69 4 41 4·87 6·03 3·02 3·25 Ilmenite 5·93 5 17 5·49 4 71 6·69 6·23 0·76 1·37 Apatite 1 68 1·68 1·79 0·67 2·02 2·02 1·01 0·60 Calcite (5 90) III.″ (III)IV (III)IV. ·IV. III III. I. (II.) I. Symbol 5″ (1) 2 (1) 2 (1) 2 5 5 5 5 3 (1) 2 (1) 2 2 ·3 3 1 2 4 2 2 2 4 4 4 3 So far as the basalts are concerned, Daly (1924) has published several analyses by Washington of rocks which are closely similar to those of corresponding rocks of Western Samoa. It is also interesting to find borne out by analysis the close petrographic similarity between Samoan and Hawaiian rocks already remarked modally by several earlier writers. Daly has also described several trachytes from Eastern Samoa, but none appears to be analogous to that from Masina. He has discussed (1924) in lucid manner the association in Samoa and elsewhere of basaltic and trachytic types, tracing the descent of these and of rocks intermediate between them, by the process of differentiation from the magma of a general basaltic substratum, which at times may, however, have formed syntectics with crust rocks. The relations in time and space between the basalts and the trachyte collected by Dr. Thomson and Mr. Wood are unknown, but they are likely to be those established for American Samoa by Daly (1924), so that this latter writer's conclusions have a direct bearing upon the discussion of the genesis of the various rocks described in this paper. It may be mentioned, however, that Washington (1923, pp. 365–6) in reviewing the evidence afforded by similar association of basalt and trachyte in Hawaii in the light of a complete series of analyses, concludes that it is adverse to Daly's (1911) hypothesis that the trachytes in those islands have originated by assimilation of limestone by basaltic magma.

Literature of West Samoan Petrography. Thomson (1921) furnishes a summary of the types of igneous rock then known from West Samoa and the full bibliography attached to his paper has proved very useful to the present writer. Many of the papers listed there have petrographic interest. Many, however, have not been accessible locally, and the following list embraces only those examined by the writer that contain important petrographic material; it further includes a few papers not directly concerned with Samoan petrography which have been referred to in the text. These last are distinguished by an asterisk immediately following the date. Cross, W., 1915.* Lavas of Hawaii and their Relations, U.S. Geol. Surv. Prof. Paper No. 88. Daly, R. A., 1924. The Geology of American Samoa, Pub. No. 340, Carnegie Inst. of Washington. —–1911* Magmatic Differentiation in Hawaii, Journ. Geol., vol. 19, p. 309. Jensen, H. I., 1907. The Geology of Samoa, and the Eruptions in Savaii, Proc. Linn. Soc. N.S.W., vol. 31, pp. 641–72. Kaiser, E., 1904. Beitraege zur Petrographie und Geologie der Deutschen Suedsee-Inseln, Jahrb. d. K. Preuss. Geol. Landes. u. Bergak., Bd. 24, H. 1, pp. 91–121. Klautzcsh, A., 1907. Der jeungste Vulcanausbruch auf Savaii, Samoa, Jahrb. d. K. Preuss. Geol. Landes. u. Bergak., Bd. 28, H. 2, pp. 169–182. Moehle, F., 1901. Beitrag zur Petrographie der Sandwich—und Samoa—Inseln, Klett & Hartmann, Stuttgart. Republished in 1902; (see Thomson, 1921.) Skeats, E. W., and Summers, H. S., 1912*. The Geology and Petrology of the Macedon District, Geol. Surv. Victoria Bull. No. 24. Thomson, J. A., 1921. The Geology of Western Samoa. N.Z. Journ. Sc. & Techn., vol. 4, pp. 49–66. Washington, H. S., 1923.* Petrology of the Hawaiian Islands, Am. Journ. Sc., vol. 5, pp. 465–502; vol. 6, pp. 100–126, 338–367, 409–423. Weber, M., 1909. Zur Petrographie der Samoa-Inseln, Abhandl. K. Bayer. Akad. d. Wiss., 2 Kl., Bd. 24, Abth. 2, pp. 290–310.