Oceanic Islands: Summary

The compositions of the volcanic oceanic islands are highly variable, some being quartz saturated and tholeiitic with compositions ranging from picrites to basalts, icelandites and rhyodacites. Others are mildly alkaline with alkali basalts to mugearite, hawaiite, benmoreite, trachytes, others are basanites-tephrites-phonolites, and nephelinites and melilite mela nephelinites such as the Bermudites also occur. Recently carbonatites have been found on both the Canary and the Cape Verde Islands. However, as with the oceanic basalts, most element pairs have definite and sharp boundaries. One bound seems to be the most tholeiitic basalt. Possibly the other is the most alkaline basanite but this is a matter which needs a good deal of investigation yet, (August, 2004)

A wide and continuous range of soda and potash is found. While the most potassic islands such as those of Jan Mayen, Tristan da Cunha, Gough and Heard have a Na/K ratio of about 1, in others in may be as high as 4. Na/K decreases in a curve with fractionation and separating the more potassic from the more sodic is not easy.

Below we show a diagram for over 8000 of the more sodic OIB's, the four groups mentioned above being excluded. Even so the soda ranges from 4.5% in trachytes to 12% in some phonolites, while K ranges from 3 to 7% in the same rocks. A few tholeites have Zr/Nb of 10 - 30, (the main tholeiitic centres of Iceland, Hawaii, Galapagos and Reunion being omitted) but the bulk of the rocks lie between 2 and 10.5. La/Sm shows a great range from 5 - 500. Lerzolites and harzburgites and other xenoliths have been omitted as have the carbonatites.

	Huge variation diagram of the more sodic OIB's. This is rather dominated by the Canary islands which make up a quarter of the total. Note the strong trend towards pantellerite, not seen in the more potassic rocks. About 40,000 data points are shown in this diagram, our largest yet! (March,2003)
	Variation diagram of the more potassic island groups, Jan Mayen, Tristan da Cunha, Gough and Heard Islands. Other groups or islands including the Society Is and Samoa are also quite potassic though slightly less so.
	The absence of pantellerites from the potassic members may be real, though the sample is not large. There is little difference in the major elements, but K, Rb, Ba, etc are about 40% higher in the potassic group on average, as are the light REE. The HFSE elements are lower. It would be interesting to compare with a selection of the most sodic rocks, eg Westmanneyjar, Madeira, Mangaia, Tubuai, St Helena, but the number of samples is rather small for positive results. As can be seen, all alkalines rocks have similar fractionation trends but the lack of build up of Fe, Ti, V at 6% MgO contrasts strongly with both ORBs and the tholeiitic island series. The ankaramite cumulate trend contrasts strongly with the picrite or troctolite trends of tholeiites and alumina steadily increases down to about 1% MgO, whereas in tholeiites and tholeiitic ORBs it decreases. Tholeiites and alkali basalts can therefore not be shown on the same diagram.
	The range of fractionation trends for OIB's is large but never goes beyond the confines of the fractionated NMORBs of East Pacific Rise (Regelous, (1999) (red squares), and the unfractionated NMORB to EMORB primary degree-of-partial-melts trend, (Kamenetsky et al, 2000), (blue diamonds). The minimum Zr/Nb for tholeiites is 10, nephelinites etc are as low as 2.5-3 Note also that the sodic St Helena and very sodic Tubuai nephelinites-analcitites do not differ much in Zr/Nb from the potassic Tristan da Cunha. Some of the St Helena trachytes go a little wild in contrast to the regular trend of the Socorro trachytes-pantellerites. Other fractionating pairs of LILE/HFSE elements, eg La/Lu, La/Sm, Nb/Y show similar but not identical relative trends.
	MgO histogram for all OIB's:. The peak at the low MgO end (trachytes) MAY indicate a high proportion of these but may indicate selective sampling. Few people could resist picking up a trachyte amongst several cubic km of boring basalt,
	SiO2 histogram, showing a bimodal spread with maxima at alkali basalt-basanite and at trachyte-phonolite.
	TiO2 histogram. again we see a bimodal distribution with peaks at basalt-basanite and at trachyte.
	Mg/Cr,Ni,Co,Zn,Co. All OIB's: Note that Cu declines with decreasing MgO while Zn seems to peak at about 3% but builds up again in trachytes-phonolites.
	Zr/ Nb,Y,Rb,Ba. - All OIB's: Nb greatly exceeds Y while Ba is extremely high.
	Zr vs Mgo for all alkaline oceanic island basalts. Alkali basalts at 5-6% MgO, ankaramites to right, trachytes, commendites, phonolites and pantellerites to left. Y, Nb and the heavy REE show much the same distribution, (see "Global Distribution of Elements"). This is probably what Professor Louis Ahrens of Capetown was referring to many years ago when he claimed that trace elements were "Log-normally distributed".
	MgO vs FeOT - All OIB's: This has been reproduced to show the very different distribution of Fe (and Ti,V) compared to both ORB and OIB tholeiites. The cumulates to the right are ankaramites (Ol-Cpx) not troctolites (Ol-Plag) and no iron-enrichment trend (or Ti or V-enrichment trend) is present. Though some trouble has been taken to remove xenoliths and cumulates from the file, many are unlabelled and those samples of more than 15% FeOT are probably xenoliths. There has been no program to analyse glass-only samples for OIBs.
	La vs REE - All OIB's:
	All OIBs - MgO vs CaO
	< Total alkalies vs MgO, Plots both against silica and MgO (or mg#)are regarded as showing the range of refraction, but the display appears quite different as seen here and below. However the limit of about 23% MgO equates to the extreme ankaramites seen at right in the diagram, with, at left, phonolites at the top, trachytes in the middle and commendites-rhyolites below.
	< Total alkalies vs silica. While the tholeiite-alkali basalt division may be obvious petrographically, there is no separation chemically. Instead the basalts form a solid group centred on 45% silica, 4% alks. Most would be inclluded in a circle of 3% SiO2 radius, leaving quadrants of low silica-low alks (mainly ankaramites), low silica high alks, basanites; high silica, high alks, mainly hawaiites-mugearites; high silica, low alks = tholeiites.

We extend out warmest thanks to the staff at "GEOROC" database, especially to Dr Barbel Sarbas, for their help in compiling the excessively large files which are used to illustrate the compositions of not only the OIB's but other series as well.