DISCUSSION

Age of Collection Localities

The biostratigraphic ranges of outcrops studied in this project are shown in Figure 3 (after Hikuroa 2005). Detailed composite lithostratigraphic sections in Figure 4 show the stratigraphic positions of belemnite bearing localities (after Hikuroa 2005). Those of Hikuroa (R.750 collections) are considered first, and we correlate those of Quilty (Q1 etc.) with those of Hikuroa (Figure 2, Figure 3, Figure 4). The ranges of belemnite genera or families in the Latady Group are compared with those in New Zealand and Madagascar in Figure 7. Quilty used international stages for some of his localities, but retained the New Zealand Heterian Stage (middle Callovian to lower Kimmeridgian strata, Campbell 2004) for some Upper Jurassic localities he could not correlate with international stages. Where possible we use ages based on ammonites, which we consider most reliable. Otherwise we use bivalves, in particular the retroceramids, which are well constrained biostratigraphically elsewhere in Antarctica (Crame 1982), New Zealand (Stevens 1997, figure 65; Hudson 2003, table 1) and South America (Damborenea 1990, 1996), and also the buchiids Malayomaorica malayomaorica Jeletzky (Crame 1990; Stevens 1997) and M. occidentalis (Jones and Plafker 1977; Jeletzky 1983).

Outcrop R.7504. The collection consists mostly of Belemnopsis, including Belemnopsis cf. B. stevensi. Hibolithes sp. C is also present. The belemnites are not age diagnostic. Based on the presence of the ammonite Perisphinctes (Discosphinctes) cf. P. antillarum (see Quilty 1970, pl. 25, figure 14), the locality is assigned a latest early to middle Oxfordian age (Perisphinctes-Araucanites Association in von Hillebrandt et al. 1992, table 12.3).

Outcrops R.7505 and R.7506. The collections are from different outcrops but the same stratigraphic horizon and contain the same fauna. Two belemnite forms were collected, aff. Brevibelus and Hibolithes catlinensis, the latter of Bajocian to Bathonian age in New Zealand. The associated fauna includes the ammonites Stephanoceras cf. S. bigoti, Teloceras cf. T. lotharingicum, and Megasphaeroceras cf. M. rotundum (Quilty 1970), and is dated as latest early Bajocian (Humphriesianum Ammonite Zone von Hillebrandt et al. 1992, table 12.2). The belemnites represent an early occurrence of Hibolithes, not present in the four assemblages recognised by Doyle et al. (1996).

Outcrop R.7507. This outcrop has a number of localities that contain a diverse assemblage of Pachyduvalia (two species), Duvalia, Rhopaloteuthis, and several species of Belemnopsis. The associated faunas are dated as latest Bathonian – early Kimmeridgian; based on the widespread Retroceramus galoi (Boehm), Retroceramus subhaasti (Wandel) and also Neocrassina marwicki (Quilty). The stratigraphically lowest locality R.7507.1 contains Neocrassina marwicki, which ranges from latest Bathonian – mid late Kimmeridgian (Hikuroa 2005). Retroceramus galoi is recorded from Q2-4 and ranges mid-middle Callovian to early Kimmeridgian (Hudson 2003, table 1). Localities R.7507.3, 5, 6, 7 have no age diagnostic fauna, but as they lie stratigraphically between R.7507.1 and Q4 (see Figure 4), must be younger than latest Bathonian and older than mid middle Callovian. R.7507.7 lies ~40 m below R. galoi and is assigned an ?early to middle Callovian age. Retroceramus subhaasti occurs concurrently with R. galoi at Q2. In the New Zealand stratigraphic sequence the two occur together from mid early Kimmeridgian to latest early Kimmeridgian, and that age is inferred for Q2. Localities Q1 and R.7507.8 and 9 are stratigraphically above Q2, and therefore younger than latest early Kimmeridgian (Figure 4).

Outcrop R.7508. The belemnites identified at generic level are all Hibolithes, and include H. cf. H. marwicki Stevens. Presence of Malayomaorica malayomaorica indicates a late Oxfordian – early Tithonian age (Stevens 1997, figure 55). Another species of Malayomaorica, M. occidentalis (late Kimmeridgian – mid middle Tithonian), is also known from the Hauberg Mountains Formation. Malayomaorica occidentalis occurs stratigraphically above M. malayomaorica (Jones and Plafker 1977; Jeletzky 1983) restricting the upper range of the latter to late Kimmeridgian. Thus, the age of outcrop R.7508 is late Oxfordian to late Kimmeridgian.

Outcrop R.7512. Indeterminate Belemnopsis, of no age-diagnostic value, are present. Presence of Malayomaorica malayomaorica indicates a late Oxfordian to late Kimmeridgian age.

Outcrop R.7517. The belemnites have affinities with Belemnopsis stevensi. Based on the presence of Retroceramus cf. R. stehni (Hikuroa 2005), Belemnopsis cf. B. stevensi at outcrop R.7517 is assigned a latest Bathonian to latest early Callovian age (the range of R. stehni see Damborenea 1990, 1996; Hudson 2003).

Outcrop R.7519. A diverse assemblage that includes Hibolithes sp. A (R.7519.60.30), Belemnopsis sp. C (R.7519.4.3; 26.4; 26.13), Belemnopsis cf. B. stevensi (R.7519.5.20), Dicoelites sp. (R.7519.17.11; 24.51), and indeterminate Hibolithes and Belemnopsis. Based on the presence of Retroceramus cf. R. stehni the fauna is assigned a latest Bathonian to latest early Callovian age.

Outcrop R.7523. ?Belemnopsis sp. indeterminate. In the absence of any reliable age diagnostic fauna, a latest early Bajocian to late middle Bathonian age is inferred based on the range of the crinoid Chariocrinus latadiensis Eagle and Hikuroa (see Eagle and Hikuroa 2003) found in association.

Q1. This locality lies stratigraphically above R.7507.8 (Figure 4). It contains Belemnopsis sp., and the same genus forms part of the fauna at R.7507. The locality lies stratigraphically above the base of the R. subhaasti zone (see discussion for R.7507 above), indicating an early late Kimmeridgian age (Stevens 1997, figure 65).

Q2. This locality lies stratigraphically below R.7507.8 (Figure 4). It contains Hibolithes aff. arkelli, Belemnopsis sp. B (UTGD87162b), Retroceramus galoi, and R. subhaasti. The two bivalves are found concurrently in the New Zealand stratigraphic sequence from mid early Kimmeridgian to latest early Kimmeridgian (Hudson 2003). This locality is older than the middle Tithonian occurrences of H. arkelli s.s. in New Zealand.

Q4. This locality is stratigraphically higher than R.7507.7. It contains non-diagnostic Belemnopsis sp. (UTGD87199). Q4 is dated as mid middle Callovian to early Kimmeridgian as it contains R. galoi.

Q6. (= R.7507.6, see Figure 4). Dated as younger than latest Bathonian and older than middle Callovian (see above). This locality contains an indeterminate belemnite (UTGD87285a).

Q9. (= R.7504.3, see Figure 4). Dated as latest early to middle Oxfordian, based on the ammonite Perisphinctes (Discosphinctes) cf. P. antillarium Jaworski (see above). Contains non age-diagnostic Belemnopsis sp. (UTGD87366c) and an indeterminate belemnite (UTGD87377a).

Q10. (= R.7505.2 and R.7506.8, see Figure 4). Dated as latest early Bajocian based on ammonites (Humphriesianum Zone, see above). Contains aff. Brevibelus and Hibolithes catlinensis. The latter ranges Bajocian to Bathonian in New Zealand.

The remaining Quilty localities are not correlated with any Hikuroa locality.

Q13. Dated as mid middle Callovian to early Kimmeridgian on the presence of Retroceramus galoi. Contains Duvalia aff. rhopaliformis (UTGD87347a), Belemnopsis sp.G. (87356a, 87357c), Belemnopsis sp. (UTGD87349a, 87353c, 87357a), and an indeterminate belemnite (UTGD87347b). Age possibly Callovian to Oxfordian, based on Duvalia. Duvalia rhopaliformis (Combémorel 1988) is late Bathonian in Madagascar.

Q15. Contains an indeterminate belemnite (UTGD87361d). Dated as mid middle Callovian to early Kimmeridgian based on R. galoi.

Q17. Contains non age-diagnostic Belemnopsis sp. D (UTGD87058c) and Belemnopsis sp. Dated as mid middle Callovian to early Kimmeridgian based on R. galoi.

Belemnite successions in West Antarctica and South America

Doyle et al. (1996) recognised a succession of four belemnite faunas in West Antarctica (Alexander Island and eastern Antarctic Peninsula), and South America (Austral and Neuquen Basins, Argentina). They are: (1) a Brevibelus-Dicoelites fauna (Aalenian-Bathonian); (2) a Rhopaloteuthis-Belemnopsis fauna (Bathonian-Oxfordian), the Belemnopsis being "European type" forms with a markedly hastate outline and strongly depressed cross section, as opposed to the robust, cylindroconical, less depressed Gondwana type Belemnopsis; (3) a Belemnopsis fauna (Gondwana type Belemnopsis, Kimmeridgian and Tithonian); and (4) a Hibolithes fauna (Tithonian-Berriasian).

Using the ages based on associated faunas for the belemnites studied here, and including the taxa and ages of Mutterlose (1986), the belemnite succession for the Latady Group can be represented by: (1) a sparse aff. Brevibelus-Hibolithes fauna (Bajocian); (2) a fauna dominated by Belemnopsis but with Hibolithes and rare Dicoelitidae and Duvaliidae (Bathonian-Oxfordian); and (3) a Hibolithes-Belemnopsis fauna with occasional Duvalia, Pachyduvalia, and Produvalia (Kimmeridgian-Tithonian).

We have described several taxa as having affinities with New Zealand species, but in most instances some differences in morphology are evident. This is not surprising. Belemnite species can be extremely variable in morphology (cf. Jeletzky 1972; Challinor 1975), and variation in the adult is often compounded by changes in shape during ontogeny (many of our specimens are juveniles). Add to this probable intraspecific differences between regional populations (development of clines), and the sparse nature of the Ellsworth Land material, often limited to single specimens or fragments, and the difficulties are clear. Nevertheless, we are confident that the West Antarctic specimens that we have compared affinities with New Zealand taxa are similar or identical to them.

Diversity of Taxa

The large number of taxa recognised may seem unusual or excessive, but the outcrops from which they were collected span much of the Middle and Late Jurassic. The situation is not without precedent. At least 25 belemnite taxa, some undescribed, are present in the New Zealand Toarcian/Aalenian to Tithonian sequence (e.g., Stevens 1965; Challinor 1974, 1975, 1979, 1980, 1996, 1999, 2003 and unpublished data e.g., Hudson 1999). It is possible we have misinterpreted some specimens, recognising them as distinct when they should be included in another described taxon. Conversely, we may have united specimens in a single taxon when they are not conspecific. This is most likely to have happened with Belemnopsis, where the material is diverse and fragmentary, but is also possible with other genera.

Regional Assemblages

Belemnite assemblages of Latady Group, Madagascar and New Zealand are compared in Figure 7. Chronostratigraphic distributions are based on families (Dicoelitidae (Conodicoelites and Dicoelites) and Duvaliidae (Duvalia, Pachyduvalia and Rhopaloteuthis) or genera (Belemnopsis and Hibolithes). Data on Madagascan taxa are from Combémorel (1988); New Zealand from Stevens (1965), and Challinor (1979, 1980, 1999, 2003); Latady Group from Mutterlose (1986) and Hikuroa (2005).

Belemnopsis is present in all three regions. Those of Latady Group range from early Callovian to late Kimmeridgian. Two morphological groups are dominant. The first contains Belemnopsis cf. B. stevensi and Belemnopsis aff. keari (Mutterlose 1986). This group ranges from early Callovian to late Kimmeridgian, and correlates with the Belemnopsis annae, B. keari and B. stevensi assemblage of the New Zealand Heterian and early Ohauan Stages (middle Callovian to late Kimmeridgian). The second group consists of small robust taxa (Belemnopsis sp. D, E, F, G, and H) and are present in the early and middle Callovian, early and late Oxfordian, and late Kimmeridgian. Similar small robust Belemnopsis also occur in New Zealand. They are Belemnopsis rarus (early Ohauan = late Kimmeridgian, Challinor 1980, figure 24-33), B. kiwiensis (middle Heterian = early Kimmeridgian, Challinor 2003, figure 37-42), and a new undescribed taxon from the early or middle Callovian. All have short stratigraphic ranges and form a very small part of the New Zealand belemnite assemblage during Callovian to Kimmeridgian time. They are not thought to be conspecific with any Latady Group taxon.

Combémorel (1988, pl. 2-11, figure 47) described 10 species of Belemnopsis ranging from the late Bathonian to late Tithonian of Madagascar. Most are very large, some have very deep, narrow, long ventral grooves; all can be readily distinguished by several features from New Zealand and Latady Group taxa. We conclude that there are strong affinities between some New Zealand and Latady Group Belemnopsis but few between those of Latady Group and Madagascar.

Hibolithes is present in the Latady Group, New Zealand, and Madagascar. They range from c. early Bajocian to early Tithonian in Latady Group; from c. early Bajocian to middle Callovian and in the early to middle Tithonian in New Zealand; and from late Bathonian to late Tithonian in Madagascar (and continue into the Early Cretaceous). Latady Group contains Hibolithes catlinensis in the Bajocian and the same taxon is present in the Bajocian-Bathonian of New Zealand. Hibolithes sp. B (a long grooved New Zealand type taxon) occurs in the late Kimmeridgian, Hibolithes aff. arkelli in the early Kimmeridgian and early Tithonian, and Hibolithes cf. H. marwicki in the early Tithonian.

Mutterlose (1986) described Hibolithes aff. marwicki, H. aff. arkelli and Hibolithes sp. from his localities 7, 8 (Bean Peaks, c. 15 km northeast of R.7508) and Wilkins Mountains (localities 2, 4). His figured specimens (Mutterlose 1986, figures 4a, b, fig. 5a, b, e-h) are of relatively well-preserved calcareous guards. Hibolithes aff. marwicki is probably identical with marwicki s.s; H. aff. arkelli and Hibolithes sp. are morphologically close to H. arkelli s.s.

Crame et al. (2000) described a fragment identified as Hibolithes aff. arkelli, recovered from a bottom trawl in the eastern Weddell Sea. Willey (1973) described Hibolithes antarctica and H. sp. nov. (Berriasian, West Antarctica) from Alexander Island. The presence of similar taxa of Tithonian or Tithonian-Berriasian age in the region makes confident identification difficult.

Mutterlose (1986, figure 4e, f) also described Hibolithes aff. verbeeki (Kimmeridgian-Tithonian) from locality 8 (Bean Peaks). The figured specimen is strongly hastate, with a very depressed cross section and a short ventral groove. It resembles some specimens of Hibolithes flemingi Spath (Kimmeridgian-Valanginian, Madagascar, Combémorel 1988). Hibolithes sp. C (R.7504.3) is short grooved as are most Madagascan Hibolithes.

Combémorel (1988, pl. 12-20, figure 47) described six species of Hibolithes from the Bathonian-Tithonian of Madagascar (and others from the Early Cretaceous). Except for Hibolithes aff. verbeeki and Hibolithes sp. C, most Latady Group taxa differ markedly from Madagascan Hibolithes. Most of the latter are large, almost all are short or very short-grooved forms, some are strongly hastate and clublike in shape. We conclude that most Hibolithes from Latady Group resemble New Zealand taxa, and that one at least (H. catlinensis) is conspecific. Two, Hibolithes aff. verbeeki and H. sp. C, resemble Madagascan taxa.

Dicoelitidae are present in all three regions. Dicoelites and Conodicoelites are present in the Bathonian – Callovian of New Zealand (Hudson 1999, and new unpublished data) and Dicoelites briefly in the late Kimmeridgian (Challinor 1980). Dicoelites is present in the early Callovian of Latady Group. Conodicoelites is not represented in the Latady Group collections studied here, but Stevens (1967) described Conodicoelites sp. from Lyon Nunataks c. 0.5° northwest of Mount Hirman. He suggested a Kimmeridgian age, probably based on the similarities to New Zealand Conodicoelites (now known to be Bathonian and Callovian in age; Hudson 1999, and new unpublished data). Those from Lyon Nunataks are not conspecific with, but are probably of about the same age as the New Zealand taxa, and with the R.7519.17/24 Dicoelites (i.e., Bathonian/Callovian). Dicoelites occurs in the late Callovian-early Oxfordian of Madagascar but is rare (Combémorel 1988, pl.1, figure 47) and Conodicoelites is apparently absent. The presence of Conodicoelites and Dicoelites in both New Zealand and Ellsworth Land at about the same time suggests affinities between the dicoelitids of the two regions.

Duvaliidae are not known from New Zealand. They range from late Bathonian to early Kimmeridgian in Madagascar, and reappear in the late Valanginian (Combémorel 1988, pl. 21-23, figure 47). They appear in Latady Group in the late Bathonian (R.7507.1), are present somewhere in the middle Callovian – early Kimmeridgian (Locality Q13, shown arbitrarily as middle Oxfordian in Figure 7), in the late Kimmeridgian (R.7507.9), and in the late Tithonian (Produvalia aff. neyrivensis, Mutterlose 1988). The affinities of Latady Group Duvaliidae are clearly with Madagascar.

Passaloteuthid belemnites are present in Latady Group and New Zealand but have not been recorded from Madagascar. Brevibelus is present in the Toarcian-Aalenian of New Zealand, a similar age range to the genus in Europe. Aff. Brevibelus (Q10) is younger, latest early Bajocian. A taxon resembling Brevibelus but larger, is present in the lower Bathonian of New Caledonia (Challinor and Grant-Mackie 1989, the specimens described as Belemnitidae (=Passaloteuthididae) gen. et sp. nov.). Taxa resembling Brevibelus occur on Alexander Island, Antarctic Peninsula in the Aalenian-Bathonian, and in Neuquen and Mendoza provinces of Argentina during the Aalenian-Bajocian. However, the name Brevibelus is placed within quotation marks, and the stratigraphic ranges are dashed (Doyle et al. 1996, figure 2), suggesting that the data are suspect or the material has not been studied in detail. These data suggest that a taxon resembling Brevibelus was present along the southern coast of Gondwana during the Early and early Middle Jurassic. The relationship between Brevibelus of New Zealand and aff. Brevibelus of Latady Group is not clear.

Inter-Regional Assemblages

Provided the ages attributed to the taxa discussed here are correct, it is possible to suggest inter-regional migration pathways.

The Duvaliidae originated in Bathonian time (Doyle et al. 1994). They apparently migrated southwards from the central Tethys to Madagascar via an intermittent trans-Gondwana seaway formed during periods of high global sea-level, which initially developed in the early Middle Jurassic. Geophysical and other investigations into the development of that seaway are discussed below. Rhopaloteuthis appeared in the Bay of Antarctica (Figure 8, in Latady Group strata) in c. the latest Bathonian, and is present in the Bathonian-Oxfordian of Alexander Island (Doyle et al. 1996).

Hibolithes and Belemnopsis appeared in the western Tethys in the Bajocian (Riegraf 1980), and presumably followed a similar migration pathway to Madagascar as the Duvaliidae (Combémorel 1988). The origin of the New Zealand late Callovian-late Kimmeridgian Belemnopsis stevensi group is uncertain but its constituents possibly evolved from strongly grooved hastate ancestors similar to the Madagascan early Callovian Belemnopsis andranomavoensis Combémorel 1988, and B. latesulcatus (d`Orbigny 1845) (see Combémorel 1988). The appearance of Belemnopsis aff. stevensi in Ellsworth Land (early Callovian) and the B. stevensi group in New Zealand (middle Callovian) could be interpreted to indicate migration from Madagascar to Ellsworth Land to New Zealand in that order. But the times of appearance are not greatly different, and could be readily explained by accidents of collection, or by imprecision in the dating methods used by different workers.

The origin of the long-grooved Hibolithes arkelli H. marwicki group is not evident. There appear to be no possibly ancestral taxa in Madagascar. They appear in the early Kimmeridgian of Ellsworth Land (Q2, Figure 7) and disappear in the early Tithonian (M, Figure 7), i.e., earlier than their appearance/disappearance in the middle Tithonian in New Zealand (A, Figure 7). The appearance of Hibolithes arkelli and H. marwicki (and also of H. kaimangoensis Challinor 1996, and H. mangaoraensis Stevens 1965) in New Zealand has been interpreted as indicating evolution in a region somewhere adjacent to New Zealand, and subsequent entry into the New Zealand area in a series of migratory events (Challinor 1999). That region of origin is apparently the Latady Basin, or another location within or close to the Bay of Antarctica (Figure 8).

The first appearance of Dicoelitidae is in the Toarcian-Aalenian of western Arctic Canada (Jeletzky 1980), but their major development and radiation appears to have been in and from the western Tethys in c. the Bathonian-Callovian. The origin of the Ellsworth Land dicoelitids is obscure. They are rare in Madagascar and in Ellsworth Land (although that could be due in the latter region to the widely scattered outcrops, Figure 1, Figure 2) but are moderately abundant in New Zealand. A Tethys-Madagascar-Ellsworth Land-New Zealand migratory pathway is possible. Dicoelitidae are present in southern South America in the Callovian (Combémorel 1988) and Aalenian-Bajonian (Doyle et al. 1996). The possibility of a migratory path from western Arctic Canada via the west coasts of the Americas to Ellsworth Land cannot be completely excluded but seems unlikely. We know of no published records of dicoelitids in the western Americas apart from those mentioned, and a western Americas migration would be partly against prevailing ocean currents (Figure 8).

Although rich belemnite assemblages are present in eastern Indonesia (Bajocian to Neocomian, Challinor and Skwarko 1982, Challinor 1991a) and Papua New Guinea (Challinor 1990), most are morphologically distinct from New Zealand and Latady Group taxa. Most Hibolithes are short grooved; none are similar to Hibolithes marwicki or H. arkelli. Conodicoelites and Dicoelites are of distinctive morphology and all are much larger than New Zealand and Latady forms. The Belemnopsis are all larger and more strongly grooved than New Zealand and Latady taxa; Oxfordian-Tithonian members are all part of an evolving lineage (Challinor 1989). Some late Kimmeridgian and early Tithonian members of that lineage were originally described as subspecies of New Zealand late Tithonian Belemnopsis aucklandica (von Hochstetter 1863), but are now regarded as close homeomorphs of the latter. When these morphological differences are considered, an eastern circum-Gondwana migration from Indonesia to New Zealand and then Ellsworth Land also seems improbable for ancestral Latady Group and New Zealand forms, and again would be largely against prevailing ocean currents.

Mutterlose (1986) envisaged migration routes that were similar to some degree. He suggested a trans-Gondwana route for the Duvaliidae (represented by his single specimen of Produvalia) during the late Tithonian. He also suggested that Conodicoelites, Hibolithes, and Belemnopsis may have also migrated trans-Gondwana, or alternatively via an eastward circum-Gondwana route, during the Kimmeridgian and early Tithonian.