A.B. SMITH ETAL. 



Faunal diversity and water-depth. The faunas that we have studied 

 come from three very different palaeoenvironments, and conse- 

 quently have very different constituent species and overall diversities. 

 The sequence at Santander was deposited in a shallow water (i.e. ca. 

 40-100 m) storm-disrupted, open-shelf environment. It contains by 

 far the most diverse fauna (33 species), with most species coming 

 from the shallowest facies (horizons 4 and 7/8). Not only is this 

 habitat rich in species, but it is dominated by species of cassiduloid 

 and regular echinoid. There are 12 regular echinoid, 9 cassiduloid, 3 

 holectypoid, 6 holasteroid and 3 spatangoid species. 



Virtually all of the species from the carbonate platform deposits of 

 Olagazutia are also present at Santander, but at Olazagutia fewer taxa 

 have been recorded. These beds were probably deposited at a similar 

 water depth, but in a more sheltered setting, with less disturbance and 

 less clastic influx. There are just nine species from Olazagutia. This 

 may be in part a sampling artifact (smaller outcrop area), but the 

 paucity of regular echinoids and cassiduloids suggests that the low 

 diversity here may be genuine and reflect a lack of shallow-water 

 elements. 



The fauna of Santander is most similar to that described from the 

 Limburg region around Maastricht, where comparable shallow- 

 water deposits are found. 15 out of the 27 named species and 19 out 

 of the 26 genera recorded at Santander also occur in the Maastricht 

 district. The genera that appear at Santander, but which are not 

 represented in Limburg, fall into two catagories. First, there are 

 shallow-water taxa, such as Zujfardia, Clypeolampas and 

 Acanthechinus, whose distribution is primarily Tethyan, and sec- 

 ondly there are rare holasteroids (Ojfaster and Galeaster) that were 

 primarily living in deeper-water settings and presumably near the 

 shallow end of their range at Santander. 



The shelf-basinal facies in Navarra, represented by black shale 

 deposits, has, by comparison, a much less diverse fauna. These beds 

 were deposited below storm wave-base, but were probably not 

 particularly deep. They lie distal to a carbonate platform and were 

 presumably more nutrient starved. The fauna of just six species is 

 strikingly different to that of Santander, and only holasteroids and 

 spatangoids are present. Stegaster is dominant in this environment, 

 but not to the exclusion of other taxa. 



In the truly deep-water, upper continental slope deposits of the 

 Bidart region stegasterids (Stegaster and Tholaster) are the only 

 echinoids present. This fauna is found elsewhere in deep-water 

 chalks and shales from south eastern Spain, Turkey and North Africa, 

 but is very different from the shelf chalk faunas that are found in 

 northern Germany and Denmark. Stegasterids were specialist sur- 

 face phytodetritivores. 



There was thus a very clear depth control on both the diversity and 

 composition of echinoid faunas in the late Cretaceous ocean. Shal- 

 low water faunas have high diversity, with a major component of 

 bulk sediment feeding cassiduloids and algi vorous regular echinoids. 

 In environments below storm wave base atelostomates totally domi- 

 nate and in really deep-water settings only stegasterid holasteroids, 

 a group specialized as surface phytodetritus harvesters, existed. 



Extinction levels. During the last stage of the Cretaceous 37% of 

 echinoid genera became extinct (Jeffery & Smith, 1997, Smith & 

 Jeffery, 1 998). However, it is almost impossible to determine whether 

 these extinctions were synchronous and instantaneous or sporadic 

 and spread over thousands or millions of years. This is because of the 

 serendipitous nature of the fossil record and because high resolution 

 correlation is rarely possible in facies where echinoids are most 

 abundant, due to the paucity of planktonic foraminifera or other 

 biostratigraphically useful fossils. Even where we do have excellent 

 stratigraphic control and continuous sections up to the Maastrichtian 



- Danian boundary, as in the Basque region (Ward & Kennedy, 

 1993), the raw distribution pattern of taxa can be very misleading 

 (Marshall & Ward, 1996, Jablonski, 1996, 1997). This is in part due 

 to sampling artefact (e.g. the Signor-Lipps effect), whereby synchro- 

 nous extinction events can appear gradual due to backward smearing 

 of last records. However, changes in facies over time can also 

 generate misleading patterns, with local disappearance being mis- 

 taken for global extinction as organisms migrate across the shelf 

 tracking facies shifts driven by sea-level change. For example, 

 although Stegaster cotteaui appears to survive to within 40 m of the 

 K-T boundary and then go extinct, this is not necessarily a correct 

 interpretation for two reasons. Firstly, the sporadic occurrence of this 

 species in the section suggests that, with more extensive sampling 

 we might expect its true last appearance to lie closer to the K-T 

 boundary. More importantly, the sister species to S. cotteaui is 

 Sanchezaster habanensis Lambert, a species known from the Lower 

 Eocene of Cuba. Therefore, the lineage must have survived the K-T 

 boundary event somewhere in the world even though it may have 

 become locally extinct in the Biscay region. 



Extinction levels at different water depths. A total of 40 

 Maastrichtian, nine Danian and nine Thanetian species are recog- 

 nized in this work, the majority coming from shelf platform settings 

 (within fair-weather wave base). Just three species from the 

 Maastrichtian shallow water fauna are known from post- 

 Maastrichtian deposits elsewhere, while only one of six species in 

 basinal mudstones survives and none of the continental slope species 

 has a post-Maastrichtian record. A literal reading of the fossil record 

 would therefore suggest that there was 90% extinction at species 

 level. 



However, sea-level changes occurring at this time resulted in 

 major changes in the distribution of sedimentary facies across Eu- 

 rope. For example, no shallow water faunas of Danian age are known 

 from Spain, carbonate platform faunas reappearing only in the late 

 Thanetian of the Pyrenees. Clearly this creates a major sampling bias 

 that needs to be taken into consideration when attempting to assess 

 levels of species extinction across the Cretaceous-Tertiary boundary. 

 For that reason survivorship of higher clades (at generic level) is 

 likely to give a more accurate picture. Maastrichtian species that 

 belong to genera that have no post-Maastrichtian record can be 

 assumed to have become extinct at around the K-T boundary, whereas 

 those with post-Maastrichtian sister groups must clearly have passed 

 through the interval. 



Of the 25 Maastrichtian shallow-water genera reported here (see 

 Appendix), 15 are known to survive into the Tertiary (a survivorship 

 of 60%). For comparison three of the five shelf-basinal genera 

 survive (60%) and one of the two continental slope genera survive 

 (50%). Two of the three surviving genera (Cvc/a5/er and Echinocorys) 

 in the shelf basinal environments also occur in shallow- water faunas. 

 If these are discounted, then we have 57% survival (13 out of 23 

 genera) of stricdy shallow- water genera, and 25% ( 1 out of 4 genera) 

 of strictly deeper-water genera. 



Our sparse data therefore points towards extinction being rather 

 more intense in shelf-basinal and upper slope settings, although the 

 small numbers make any statisfical testing meaningless. The demise 

 of the deep-water fauna is probably linked to the plankton crash 

 known to have occurred at the end Cretaceous, since the echinoids 

 that become extinct are all specialist deposit feeders reliant on 

 phytodetritus. The extinction in shallow platform settings is more 

 difficult to assess, though here again food supply may have been 

 crucial. 



Extinction levels andpalaeolatitude. Some workers (e.g., Macleod 



