Red Devonian Trilobites with Green Eyes from Morocco and the Silicification of the Trilobite Exoskeleton

Latest Emsian (Early Devonian) sediments at the famous mud-mound- and trilobite-locality Hamar Laghdad (Tafilalt, Morocco) yielded some red-coloured remains of phacopid trilobites. Closer examination revealed that the eyes of these phacopids are often greenish in colour. EDX-analyses showed that the lenses retained their original calcitic composition, possibly greenish due to Fe- and Mn-impurities, while most of the exoskeleton was silicified. The silicified parts contain elevated concentrations of iron which causes the red colour. This phenomenon is explained by the porosity of the exoskeleton in contrast to the homogeneous and massive construction of the lenses and their Mg-content. These incompletely silicified trilobites enabled a reconstruction of the silicification process in trilobites. Their diagenetic alteration probably occurred as a result of events associated with the Cretaceous transgression.


Introduction
Trilobite eyes always attracted the attention of palaeonto− logists and biologists, partially because of their often large size compared to body size but also because of the sheer fact that such a sense organ can be preserved in such detail: Even the quality and optical properties of its lenses could be exam− ined (Clarkson and Levi−Setti 1975;Bruton and Haas 2003 and references therein;Fortey and Chatterton 2003). Espe− cially phacopid eyes were often discussed in numerous arti− cles, partially because of the abundance of this trilobite group and partially because of the large size of the eyes and their lenses, respectively, of some members of the Phacopidae (Bruton and Haas 2003 and references therein).
Because of its usually strongly mineralised cuticle, it is often well preserved and thus, the ultrastructure of the trilo− bite cuticle could be studied in detail in various groups (e.g., Teigler and Towe 1975;McAllister and Brand 1989;Daling− water et al. 1991Daling− water et al. , 1993Daling− water et al. , 1999. As we show in this article, the physical properties required for optical functionality of the eyes and their Mg−content (Lee et al. 2007) influenced the diagenesis of this part of the trilobite exoskeleton. We exam− ined a number of well−preserved phacopid trilobites from Hamar Laghdad (Tafilalt, eastern Anti−Atlas, Morocco) be− cause of their peculiar preservation. At this locality, we found several enrolled specimens of Barrandeops cf. granu− lops (Chatterton, Fortey, Brett, Gibb, and McKellar, 2006) (e.g., PIMUZ 27077), a species formerly described from the DraValley (Chatterton et al. 2006). In these specimens, the exoskeleton is red except for the lenses which are green. Im− ages of one of these extraordinary fossils (PIMUZ 27077) were used for the cover of the April 2008−edition of the Jour− nal "Geology", but without an accompanying article. In this article we want to (1) show why the eyes of these trilobites are green and (2) discuss how this phenomenon can be ex− plained.

Material and geological setting
The recently discovered, seemingly perfectly preserved pha− copid trilobites were eroded from marls from close to the Emsian/Eifelian boundary from the Moroccan locality Hamar Laghdad.

Methods
In order to learn more about the origin of this coloration, we examined the chemical composition of the red trilobites from Hamar Laghdad. For this purpose, we produced thin−sections (PIMUZ 27075, 27081; Fig. 2D-H). These thin−sections and an isolated phacopid cephalon (PIMUZ 27076) were sub− jected to a series of elemental micro analyses using a scan− ning electron microscope SEM LEO 1450 VP at the Institut für Geowissenschaften Tübingen (Fig. 3). Operating para− meters were 17 kV with a working distance of 29 mm. The chemical composition of sample points was measured using an energy−dispersive X−ray analyser EDX (Oxford Instru− ments INCA EDS 200). Measuring time was 90 seconds. These examinations clearly showed that the cuticle of the isolated cephalon (PIMUZ 27076) is silicified except for the lenses, which retained their calcitic composition ( Fig   wt. % calcium, 54.0 wt. % oxygen). On the glabella of the measured specimen, five analyses rather uniformly revealed the presence of approximately 1.9 wt. % iron (which explains the red color), 1.0 wt. % calcium, 2.0 wt. % aluminum and 0.7 wt. % potassium (Fig. 3). Not all specimens from this locality are preserved identi− cally. The degree of silicification and iron−staining is highly variable. This is reflected in exoskeleton colours ranging from greenish−yellowish (when still more or less calcitic) to red (when silicified). One enrolled specimen which was thin−sectioned (PIMUZ 27081) retained much of its original calcitic mineralogy with a clear predominance of calcium and oxygen, reflected in its more greenish to yellowish col− our. The element of iron was detected by the EDX in traces only and in some places not at all. Hassan (1978) analysed the composition of a greenish calcite from Egypt and found impurities of manganese and iron ["Mn(IV) and Fe(II)"]. Based on these findings, he concluded that the calcite was precipitated "in an environment that was more acidic and with a very low oxidation potential" (Hassan 1978: 732).

g a r t a A n t i -A t l a s A n t i -A t l a s
The EDX−analyses in specimens PIMUZ 27075 and 27076 clearly showed that all red parts of the exoskeleton are silicified while the yellowish to greenish parts such as the lenses retained the calcitic composition with low contents of iron and magnesium. The examinations of the thin−sectioned cephalons revealed that the replacement of calcite by silica commenced at the outer surface (Fig. 2G, H) of the sclerites and following pore canals (Fig. 2E, H) as well as the bend in the growth lines of the intralensar sclera (Fig. 2D, F; cf. Bruton and Haas 2003). More or less simultaneously, man− ganese and iron minerals were deposited from the ventral side (i.e., from within the enrolled trilobite) in the pore canals ( Fig. 2E, H). The sediment matrix visible in the thin−sections (Fig. 2D-H) is partially silicified (red areas) and partially carbonatic (light coloured under incident light). The lenses locally are well preserved with the primary small crystal size while adjacent lenses can be already recrystallised, forming a single or few crystals (Fig. 2C, D, F).

Remarks on the local diagenesis
We found evidence for two diagenetic processes which have occurred directly next to each other at the "red cliff" (Hamar Laghdad): Dolomitisation.-A thin section of red−coloured Late Eife− lian limestones from the top of the "red cliff" is composed of relatively large, unmistakable dolomite crystals (< 2 mm). The Late Eifelian rocks contain also dolomitised fossils such as various cephalopods. Dolomitisation occurs abundantly in the carbonates of the Anti−Atlas which characteristically have a reddish colour due to iron compounds (compare Kaufmann 1997Kaufmann , 1998. At Hamar Laghdad, this reddish facies is largely restricted to the more elevated parts of the Hamar Laghdad ridge (Fig. 4).
Silicification.-The mineralogy of both the trilobite exo− skeletons and their sediment matrix indicate a local and par− tial silicification. Silicified trilobites occur (usually in associ− ation with silicified ostracods) also in other localities of the Anti−Atlas such as in the Taouz− and Ouidane Chebbi regions in roughly the same stratigraphic position. As far as we know, Hamar Laghdad is the only locality where the silici− fied trilobite exoskeletons carry the red colour.
In the Tafilalt, dolomitised rocks are often reddish in col− our, certainly due to their iron compounds−content. These dolomites occur mainly in two situations: near faults (Bernd Kaufmann, personal communication 2008;Graz and Kauf− mann 1997) and close to the erosional surface at the base of the Cretaceous transgression. To our knowledge, these dia− genetic phenomena have not been examined in detail in the study area yet and thus, a full sedimentological interpretation of these diagenetic processes is not available yet.
At Hamar Laghdad, the red facies is at least partially re− lated to the erosional processes prior and during the Creta− ceous transgression because it occurs predominantly in ele− vated parts of the Hamar Laghdad−ridge, i.e., close to the ero− sional surfaces, and it occurs also in other parts of the Tafilalt where the Devonian sediments reach close to the Cretaceous transgression surface (e.g., Ouidane Chebbi, Achguig). At the red cliff, some red−coloured erosional remnants of the Creta− ceous transgression conglomerate containing coarse silici− clastics are preserved. We suggest, tentatively, that at least the dolomitisation but potentially also the silicification at the "red cliff" were caused by the penetration of silicium−, aluminium− and iron−bearing fluids after the uplift, during the later erosion of the Devonian sediments and during the Cretaceous trans− gression. The silicification is not surprising since the Creta− ceous transgressive sediments contain sandstones, conglomer− ates (both with siliceous components and sometimes with a si− licified matrix) and thin chert−layers. The locally limited oc− currence of this preservation is not surprising since the Late Emsian to earliest Eifelian marls are less resistant towards ero− sion than, e.g., the massive mud−mound−limestones and thus, sediments of this stratigraphic intervals which were altered by transgression−related diagenetic phenomena are rare.
We also considered that the silica might have originated from the local hydrothermal activity (cf. Belka 1998;Aitken et al. 2002). This appears unlikely, however, because this phenomenon does not occur everywhere in Hamar Laghdad but only on the red cliff while hydrothermal activities proba− bly occurred in the entire Hamar Laghdad region.

Trilobite exoskeleton diagenesis
The above described patterns of the selective and partial sili− cification of the trilobite exoskeleton were probably caused by differences in porosity (pore canals in most of the exo− skeleton in contrast to the homogeneous lenses) and compo− sition (primary magnesium content of the lenses; see Lee et al. 2007). For the silicification of the trilobites, the following sequence of events is evident from the thin−sectioned speci− mens: (1) Silicification began at the outer surface of the exo− skeleton as well as in pore canals and along the bend of growth lines within the intralensar sclera.
(2) The exoskeleton became fully silicified including the intralensar sclera while the lenses stayed calcitic but may have become recrystallised.

Conclusions
Phacopid trilobites from the latest Emsian of Hamar Laghdad, Morocco are silicified to varying degrees. This fact allows a reconstruction of details of the silicification process. Silicifica− tion started at the outer surface of the exoskeleton as well as in pore canals and along the bend of growth lines within the intralensar sclera. From these silicification centres, this pro− cess continued until only the lenses remained calcitic. Addi− tional to the homogeneity of the calcite crystals of the lenses, needed for their optimal optical functionality, and their higher magnesium−content (Lee et al. 2007), the absence of pore−ca− nals in the lenses accounts for the fact that the eyes retained the greenish calcite (this colour is also encountered in various other invertebrate fossils from the red cliff as well as brachio− pod shells such as the thick−shelled Middle Devonian Devono− gypa and Ivdelinia from the Maïder Basin; Kaufmann 1998) while the rest of the trilobite exoskeleton is silicified and red because of a sufficiently high content of iron compounds. The greenish colour of the lenses might have been caused by man− ganese and iron−impurities which could not be detected with certainty with the EDX because the content was too low. This differential replacement also explains why the lenses are slightly corroded while the rest of the exoskeleton is rather well preserved. So far, this differential replacement of calcite is only known from the eyes of phacopids. Other trilobite taxa from the same strata have holochroal eyes and thus much smaller lenses which were probably more rapidly replaced by silica. The lenses of these holochroal eyes are preserved in the red, iron−stained silica exactly like the exoskeletons. The sili− cification probably happened in relation to the Cretaceous transgression because (1) the silicification occurs at the only one place at Hamar Laghdad, the red cliff, where the latest Emsian marls reach close to the transgression surface (they are more deeply eroded otherwise because of their higher clay content and thus lower erosion−resistance), (2) the Cretceous sediments display the same colour, (3) the Cretaceous sedi− ments are rich in silica (siliciclastic conglomerates and cherts), (4) this phenomenon does not occur in other parts of Hamar Laghdad in the same stratigraphic position, and (5) silicified arthropods from other localities in the Anti−Atlas do not show this coloration.
Future studies should continue by analysing the alteration of sediments below the Cretaceous transgression and analy− ses of the green colour in brachiopods and trilobite eyes as well as other fossils from the red cliff.