Strontium isotope systematics of conodonts: Middle Devonian, Eifel Mountains, Germany

Abstract

The Eifel Mountains lithological profile, the stratotype for the Middle Devonian, was densely sampled for conodonts to delineate the Sr isotopic evolution of coeval sea water. The ⁸⁷Sr/⁸⁶Sr curve obtained shows oscillations around a mean of about 0.7078 that have a frequency in the 10⁵–10⁶ yr range and an amplitude in the 10⁻⁵–10⁻⁴ range. These oscillations mimic those obtained on coeval brachiopods (Diener et al., 1996), but are shifted into more radiogenic domains by up to 10⁻⁴. Model calculations suggest that the shift is due to partial (30–40%) early diagenetic equilibration of conodonts with their enclosing bulk-rock matrix. The reality of partial recrystallisation is supported by visible neoformed blocky apatite crystals on the surface of the conodonts, despite their otherwise apparently excellent preservation, with CAI values of 1.5–2.5. We therefore suggest that studies anticipating use of conodonts as a recorder of paleoceanographic events should be confined to specimens embedded in a relatively `clean' matrix, such as pure limestone. Even these results should be cross-checked by comparison with data on coeval low-Mg shells, such as brachiopods or belemnites.

Introduction

The goal of the present study, and of similar efforts elsewhere, is to develop a high resolution chemostratigraphy for the Paleozoic. This task is considerably more difficult than was the case, for example, for the Cenozoic, due to potentially higher diagenetic overprint and to poorer stratigraphic resolution. In addition, the availability of reliable materials that record the properties of sea water is limited. The principal samples used for delineation of the details of Sr isotopic composition of Paleozoic sea water are low-Mg calcitic shells of brachiopods (Popp et al., 1986; Brand, 1991; Gruszczynski et al., 1992; Diener et al., 1996). However, the stratigraphic resolution of brachiopods in the Paleozoic is mostly inferior to that of conodonts. The apparent advantages of the latter include the fact that they have high Sr contents (Pietzner et al., 1968; Wright et al., 1984, Wright et al., 1990; Kürschner et al., 1992; Bertram et al., 1992) and that their degree of post-depositional alteration can be reliably calibrated via the colour alteration index (CAI) (Epstein et al., 1977; Königshof, 1991). For these reasons, conodonts have been viewed as potentially the best material for studies of the Sr isotopic composition of Paleozoic sea water (Kovach, 1980, Kovach, 1981, Kovach, 1984; Wright et al., 1984; Keto and Jacobsen, 1987; Kovach and Miller, 1988), providing their CAI does not exceed a value of 2.5 (Bertram et al., 1992).

These presumed advantages are somewhat mitigated by the fact that conodonts are the remains of an extinct group of chordates (Aldridge et al., 1986; Conway-Morris, 1989) and Paleozoic samples cannot be directly compared to any extant counterparts. The average composition of conodont apatite, according to Pietzner et al. (1968), can be expressed as:

$$\text{Ca}{}_5\text{Na}{}_{0.14}\text{(PO}{}_4\text{)}{}_{3.01}\text{(CO}{}_3\text{)}{}_{0.16}\text{F}{}_{0.73}\text{(H}{}_2\text{O)}{}_{0.85}$$

The OH⁻ or H₂O are not present in the lattice position, as in hydroxy-apatite, but substitute, together with carbonate, for phosphate. This is typical of francolite. Conodonts contain three chemically and morphologically different domains. The first shows in thin section a lamellar texture and contains enhanced contents of CO²⁻₃ and a higher proportion of organic matter. The second domain, the `white matter', has no lamellae, less organic matter and less carbonate. This material is more frequent in the phylogenetically older, particularly Ordovician, conodonts and probably represents recrystallised portions of conodonts. The timing of such presumed recrystallisation, pre- or post-mortem, is open to discussion (Lindström and Ziegler, 1971). The third domain is the so-called basal filling of conodonts. In well preserved conodonts, it is frequently found attached to the conodont element and is composed of small apatite crystals which are enriched in REE and depleted in Sr (Pietzner et al., 1968).

In terms of trace elements, of particular importance for the present study is the behaviour of strontium. In sedimentary phosphates, Sr substitutes for Ca in the crystal lattice (McClellan, 1980; McArthur, 1985) and its concentration in francolite should be proportional to that in a parent solution (LeGeros et al., 1980; McArthur, 1985). If the mineralization of conodont francolite has not changed in the course of geologic history, its chemistry may have been comparable to that of recent fish teeth. The latter have a Sr content of 1–100 ppb (Darlet, 1984; Elderfield and Pagett, 1986). However, the Sr contents increase very rapidly, to 1000 ppm, during very early diagenesis (Schmitz et al., 1991). Fossil fish teeth have even higher Sr contents of 1500–5000 ppm (Staudigel et al., 1985; Grandjean et al., 1987; Schmitz et al., 1991), with the variability possibly related to differences in diagenetic alteration. Similarly, well preserved conodonts contain about 3000–7000 ppm Sr (Pietzner et al., 1968; Wright et al., 1984; Bertram et al., 1992; Kürschner et al., 1992; Holmden et al., 1994).

From studies of REE in fish teeth (Bernat, 1975; Elderfield and Pagett, 1986; Grandjean et al., 1987, Grandjean et al., 1988; Grandjean and Albarède, 1989), which show a similar post-mortal enrichment as Sr, it appears that these elements are scavenged from Fe–Mn oxyhydrates, pellets and organic debris that are decomposing on the ocean floor, producing reducing microenvironments and migration of such elements into francolite. This sequence of events is believed to be the reason for preservation of a marine Sr isotope signal, since the microenvironment is buffered by marine products and marine pore waters in terms of its ⁸⁷Sr/⁸⁶Sr ratio. Subsequently, diagenetic overprint at higher pressures and temperatures results in progressive transformation of the metastable francolite into fluorapatite, which is accompanied by the loss of Na, Sr and REE (McArthur, 1985; Burnett, 1988). Our studies of conodonts with CAI 1–5 (Kürschner et al., 1992) show an average Sr decline from 3500 to 2000 ppm, although the scatter is considerable. This scatter is probably due to the variable proportions of the `white matter' in the samples, probably a primary or early diagenetically formed fluorapatite. Its higher stability may result in greater retention of early diagenetic (primary?) chemistry at comparable P–T overprint. Ramiform conodonts usually contain more `white matter' and appear, therefore, more suitable for isotopic studies. As an additional precaution, when conodont elements had to be pooled for Sr isotope measurements, we combined two or three monogeneric types in order to minimize effects resulting from the variable content of the `white substance' in the samples. Weathering also appears to result in a loss of trace elements such as Sr and Na, as demonstrated for sedimentary phosphates by McArthur (1978)and McClellan (1980). Analyses by Shemesh (1990)on recent marine apatites and onshore phosphates revealed a decrease in Sr with increasing phosphate crystallinity. The processes of enrichment and depletion of trace elements may follow the same pattern, but not necessarily the magnitude, of trace element repartitioning typical for carbonates (cf. Veizer, 1983a,Veizer, 1983b).

According to Pietzner et al. (1968)and Wright et al. (1990), exceptionally well preserved conodonts contain about 1.8% carbonate. X-ray diffraction analysis of one conodont sample with a CAI of 3 (Kürschner et al., 1992) suggests that its carbonate content is only 0.8%. This would indicate that, even at this stage, conodonts suffer from diagenetic recrystallisation.

The complex post-depositional history of conodont elements, outlined above, casts sufficient doubt as to the preservation of the original Sr isotopic signature in phosphatic remains of pre-Mesozoic age to warrant more detailed study.