1
“Hyopsodontid” affinities to elephant-shrews and northern Palaeocene afrothere origins
(Supplementary Information)
Shawn P. Zack*, Tonya A. Penkrot*, Jonathan I. Bloch†, and Kenneth D. Rose*
*Center for Functional Anatomy and Evolution, The Johns Hopkins University School of Medicine, 1830 E. Monument Street, Baltimore, Maryland 21205
†Florida Museum of Natural History, Dickinson Hall, Museum Rd. & Newell Dr., PO Box 117800, University of Florida, Gainesville, Florida 32611-7800
Contents
1. Principal Specimens Examined
2. Extant Comparative Material Examined
3. Modifications to Meng et al. (2003) and Asher et al. (2003) character-taxon matrices
3a. Meng et al. (2003)
3b. Asher et al. (2003)
4. Characters and character-taxon matrix for the new ungulate analysis
5. Analyses performed and results.
5a. Menget al. (2003).
5b. Ungulate analysis
5c. Asher et al. (2003).
6. Supplementary Bibliography
1. Principal Specimens Examined
For locality information, see Bown et al. (1994), Bloch and Bowen (2001), and Silcox and Rose (2001).
Apheliscus chydaeus
UW loc. W-22
USNM 525597: associated left P3-M3; right P3-4, M2-3; left p2-4, m2-3; left astragalus, calcaneum, cuboid, distal femur, proximal radius, proximal humerus; right humeral shaft, ulnar fragment, proximal tibia; fragments of both tibial shafts; vertebrae
UW loc. W-16A (Banjo Quarry)
USNM 525594: left astragalus
Apheliscus insidiosus
USGS loc. D-1350Q
USNM 488325: left proximal femur and proximal tibia; right distal femur (unassociated but possibly from a single individual based on degree of epiphysial fusion)
USGS loc. D-1460Q (Rose Quarry)
USNM 493903: unassociated left proximal tibia and distal tibia-fibula; right cuboid and distal humerus. USNM 521789: right calcaneum. USNM 521790 and 521791: right astragali. USNM 525593: unassociated right calcaneum, proximal femur, and distal humerus.
USGS loc. D-2035Q (Dorsey Quarry)
USNM 491971: left distal tibia-fibula. USNM 495051: right tibia-fibula. USNM 488326: left femur. USNM 525591: unassociated left astragalus and calcaneum. USNM 525592: right astragalus; USNM 525646: associated right proximal tibia and fibula.
Bighorn Basin (locality unknown)
AMNH 15696: associated crushed cranium with left P4-M3, right P4, M2-3; left mandible with p4-m3; right mandible withp4-m1, m3; left humeral shaft, and proximal ulna; right distal humerus; pelvic and vertebral fragments
Apheliscussp.
D-2037Q (McNeill Quarry)
USNM 493819: unassociated left distal humerus, distal femur, and proximal tibia, right distal femur and calcaneum
Haplomylus speirianus
UM locality SC-4 (8abc limestone)
USNM 513057-513062: proximal femora. USNM 513140: femur. USNM 513173-513175: distal femora. USNM 513239: distal tibia-fibula. USNM 513245-513247: proximal tibiae. USNM 513512: humerus. USNM 513555-513557: distal humeri. USNM 513632-513635: astragali. USNM 513655-513665: calcanei. USNM 513668: cuboid. USNM 513868: tibia-fibula.
USGS loc. D-1223
USNM 488321: left astragalus
UW loc. W-37
USNM 493902: right astragalus
UW loc. W-44
USNM 488327: two unassociated left calcanei
UW loc. W-46
USNM 488328: left calcaneum
UW loc. W-86
USNM 493901: left calcaneum. USNM 525595: unassociated left and right calcanei
Anthill across from UW loc. W-86
USNM 488329: right calcaneum
YPM loc. Y-327
USNM 525596: right astragalus
2. Extant Comparative Material Examined
Abbreviation: USNM:United States National Museum, Smithsonian Institution, Washington, DC.
Hypsiprymnodon moschatus: USNM 238443, 238444.
Rhynchocyon cirnei: USNM 537657.
Petrodromus tetradactylus: USNM 521009.
Elephantulus rufescens: USNM 283463.
Echinosorex gymnurus: USNM 448861.
Hemiechinus auritus: USNM 396508; KDR personal collection uncataloged.
Erinaceus europaeus: USNM 251764, 251765.
Dolichotis patagonum: USNM 175890.
Jaculus jaculus: USNM 308400, 477276.
Pedetes capensis: USNM 344334, 295258.
Dipodomys deserti deserti: USNM 034369, 034370.
Dipodomys ordi palmeri: USNM 05372.
Ochotona rufescens vizier: USNM 326747, 326748, 326750.
Ochotona alpina argentata: USNM 240727.
Sylvilagus sp.: KDR personal collection L1, L2.
3. Modifications to Meng et al. (2003) and Asher et al. (2003) character-taxon matrices
3a. Meng et al. (2003)
In the process of scoring Apheliscus and Haplomylus, we also re-examined Meng et al.’s (2003) codings of the extant macroscelideans Petrodromus and Rhynchocyon, which resulted in changes to the original codings. Meng et al. (2003) coded Petrodromus as retaining an astragalar foramen (Character 210). Our observations of Petrodromus astragali indicate that an astragalar foramen is lacking in this genus. Accordingly, we recoded Petrodromus as “1” for Character 210. Meng et al. (2003) also coded eight tarsal characters as missing data in Rhynchocyon, presumably because they lacked access to a disarticulated tarsus. We therefore scored Rhynchocyon for these characters (205-10, 213-4). We include the new codings for Rhynchocyon, Apheliscus and Haplomylus below.
Meng et al. (2003) ordered all characters in their ordered analysis. However, only 44 out of 87 characters with three or more states form plausibly ordered transformation series, at least given the order of states presented. In our ordered analysis only these characters are ordered.
Ordered: 4, 11, 13, 18, 23, 25, 33, 39, 42, 44, 45, 50, 53, 54, 73, 75, 77, 82, 85, 86, 91, 102, 108, 113, 119, 122-6, 139, 142, 145, 148, 152, 161, 163-5, 169, 171, 176, 195, 199.
Codings for new and modified taxa (A=0/1; B=2/3):
Rhynchocyon
0??011001?000000001110001100000000403020020021110000300130001041000A01 0111000000 0030100200 0000100010 000010002? 001B010100 0200000000 0000000000 0000011100 1000000110 0111120001 0001110?10 1101011001 0101111121 1201101111 1001111320 0010000
Apheliscus
0???1?00??0????0???110011100001000200000100000000000210012001100000000 0010002001 000010000? 00?0??00?? ?????????? ?????????? ?200000??0 ?????????? ????00???? ?????????? ?????????? ????????10 10??11???? 1011011020 1201001111 0001111300 001????
Haplomylus
??????????????????????????0?0??000401010010000000002110011001101000010 0010001201 0000200?00 00001????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ????????10 11???????? 111101102? 1201101111 0001111320 001????
3b. Asher et al. (2003)
Because of the relative novelty of the Afrotheria hypothesis, there is no available morphological analysis that samples all afrothere clades but does not sample the remainder of Eutheria. However, the character-taxon matrix of Asher et al. (2003) includes representatives of all extant afrothere clades and offers a reasonable base into which Apheliscus and Haplomylus were added. Because a full assessment of the position of Macroscelidea is beyond the scope of this study, we culled Asher et al.’s (2003) taxonomic sample to include only extant members of afrotherian groups and extinct taxa with potential afrotherian affinities. We also eliminated one outgroup (Macropus) and used Ukhaatherium as a second outgroup. Of the extinct taxa included in Asher et al.’s (2003) analysis, only Zalambdalestes was excluded as it showed no affinity with any afrothere groups in Asher et al.’s published results and its precociously specialized morphology (e.g. Kielan-Jaworowska, 1978; Fostowicz-Frelik and Kielan-Jaworowska, 2002; Wible et al., 2004) makes it a poor choice as an outgroup.
In addition to scoring Apheliscus and Haplomylus, a few additional modifications were made to Asher et al.’s (2003) character-taxon matrix. Five characters (83, 99, 110, 164, 167) were identified in which the character descriptions and scorings do not match up due to inversion of the numbers assigned to character states. In the character descriptions of Asher et al. (2003), “0” and “1” should be reversed for these characters. Additionally, Trichechus was recoded as 1 for character 118 to match the description of this character which specifically refers to “sea cows and desmostylians” as showing state 1. Loxodonta was recoded as 2 for character 141 to align it with the description of this character. Finally, Elephantulus and Macroscelides were recoded as 0 for this character to reflect the fact that these taxa retain a small coronoid process of the ulna that articulates with the anterior aspect of the humeral trochlea.
Codings for new taxa:
Apheliscus
??????????????????????????????????????????????????????0??????0???0???? ?????????? 001??1???? ???00??000 0000010001 0000000000 0????????? ????100000 0????????? ?????????? 0000100010 1100100??? ?????????? ??????
Haplomylus
??????????????????????????????????????????????????????0??????0???????? ?????????? ?????????? ???0????0? 0000110001 0010000000 0????????? ????110100 ?????????? ?????????? 00001?0010 1100100??? ?????????? ??????
4. Characters and character-taxon matrix for the new ungulate analysis
There are two existing character taxon matrices that sample a diversity of basal members of Ungulata. The first of these, presented by Thewissen and Domning (1992) does not include any of the taxa of interest, including extant Macroscelidea. This is particularly troublesome, as it indicates that characters that would resolve the position of macroscelideans within Ungulata are not necessarily included. A more recent matrix presented by Tabuce et al. (2001) densely samples both Macroscelidea and “Hyopsodontidae.” Unfortunately, their character sample is limited to the dentition.
Because of these deficiencies in existing matrices, we constructed a new character-taxon matrix, focusing on condylarths and early members of extant ungulate orders. A total of sixteen taxa were included, including one outgroup (Protungulatum), two arctocyonids, one mesonychid, one periptychid, one mioclaenid, two phenacodontids, four “hyopsodontids,” and one representative each of the orders Macroscelidea, Hyracoidea, Perissodactyla, and Artiodactyla. Taxa were scored for 52 characters sampling the cheek dentition (27) and proximal tarsus (25). Because tarsals of Eocene African macroscelideans are unknown, the OTU for Macroscelidea is a composite of Chambius (dentition) and Rhynchocyon (tarsus). Similarly, because calcanei of Eocene hyracoids have not been illustrated, we scored extant Procavia for calcaneal characters. Otherwise, the hyracoid OTU was scored based on Seggeurius (dentition), Bunohyrax (dentition), Antilohyrax (astragalus), and the unidentified hyracoids whose astragali were illustrated by Rasmussen et al. (1990). Finally, calcaneal characters for the Dissacus OTU are a composite of Pachyaena and Mesonyx, as calcanei of Dissacus remain unknown. Five characters form plausibly ordered transformation series.
Institutional Abbreviations
AMNH: American Museum of Natural History, New York City, New York; CMNH: Carnegie Museum of Natural History, Pittsburgh, Pennsylvania; DPC: Duke University Primate Center, Durham, North Carolina; IVPP: Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing; MNHN: Muséum National d’Histoire Naturelle, Paris; NMMNH: New Mexico Museum of Natural History, Albuquerque, New Mexico; SPSM: Science Museum of Minnesota (St. Paul Science Museum), St. Paul, Minnesota; UCMP: University of California Museum of Paleontology, Berkeley, California; UM: University of Michigan Museum of Paleontology, Ann Arbor, Michigan; UNM: University of New Mexico, Albuquerque, New Mexico; USGS: United States Geological Survey, Denver, Colorado; USNM: United States National Museum, Smithsonian Institution, Washington, DC; YPM: Yale Peabody Museum, Yale University, New Haven, Connecticut; YPM:PU: Princeton University Collection, Yale Peabody Museum, Yale University, New Haven, Connecticut
Specimens Examined
Protungulatum
P. donnae: SPSM 62-2028; AMNH 96357, uncataloged tarsals
Hyopsodus
H. loomisi: UCMP 44781 and uncataloged USNM specimens
H. paulus: Gazin (1968); Godinot et al. (1996)
H. sp.: USNM 493782, 495492, 521781; USGS 4725; YPM uncataloged skeleton
Apheliscus spp. (See material examined for postcranial specimens. Only additional specimens used to code dental characters are listed)
A. chydaeus: USNM 525597
A. insidiosus: USGS 12608; USNM 28328, 494986
A. sp.: USNM 494896
Haplomylus spp. (See material examined for postcranial specimens. Only additional specimens used to code dental characters are listed)
H. simpsoni: UM 65249
H. speirianus: USNM 488319, 493936, 512563, 521645
Paschatherium spp.
P. dolloi: Godinot et al. (1996)
P. marianae: Estravís and Russell (1992)
Mithrandir
M. gilianus: AMNH 16461; NMMNH P-3083
Chriacus
C. gallinae: USGS 2353, 15404; YPM 38565; AMNH 48006
Ectocion
E. osbornianus: USNM 487874, 487875, 494919; 494921, uncataloged tarsals
E. parvus: USGS 38074
Tetraclaenodon
T. puercensis: Matthew (1937); Radinsky (1966); UNM B-1168, B-897; USNM 409375
Claenodon spp.
C. ferox: AMNH 3268, 16543
C. procyonides: AMNH 16554
C. sp.: YPM:PU 17301, 17424, 17545, 18701
Diacodexis spp.
D. secans: USNM 511096, 511315, 521421, uncataloged tarsals; USGS 2352, 25383, 25282
“Hyracotherium” spp.
Sifrippus sandrae: UM 79889, 83473, 83567, 83615
“Hyracotherium” sp.: USGS 5901, 6110, 30839
Molinodus
M. suarezi: Muizon et al., (1998); Muizon and Cifelli (2000)
Dissacus
D. navajovius: AMNH 3356, 3360, 15966
D. europaeus: MNHN BR 21L
D. sp.: IVPP 5478, 5479
Pachyaena ossifraga: USGS 25292
Mesonyx sp.: CMNH 47921 (calcaneum only)
Macroscelidea
Chambius kasserinensis: Hartenberger (1986); Tabuce et al. (2001)
Rhynchocyon cirnei: USNM 537657
Petrodromus tetradactylus: USNM 521009
Hyracoidea
Bunohyrax spp.: Matsumoto (1926)
Seggeurius amourensis: Court and Mahboubi (1993)
Antilohyrax pectidens: DPC 7723a
Procavia capensis: K.D.R. pers. coll. H-1
Pliohyracidae indet.: Rasmussen et al. (1990)
Character List
1. Lower molar cusp relief
(0) Strong, cusps relatively acute
(1) Weak
2. Relative heights of molar trigonids and talonids (Ordered)
(0) Trigonids much higher than talonids
(1) Trigonids somewhat higher than talonids
(2) Trigonids and talonids subequal in height
3. p4 paraconid and metaconid
(0) Both cusps weak or absent
(1) Both cusps well developed
(2) Metaconid well developed, paraconid weak or absent
4. p4 protoconid shape
(0) Conical, p4 trigonid relatively closed
(1) Elongate, p4 trigonid relatively open
5. p4 talonid complexity
(0) p4 talonid simple, with a single well developed cusp
(1) p4 talonid complex with multiple cusps
6. Relative widths of m1-2 trigonids and talonids
(0) Talonids broader than trigonids
(1) Trigonids and talonids subequal in width or trigonids broader
7. Orientation of m1-3 trigonid cusps in buccal view
(0) Cusps erect
(1) Cusps canted mesially
8. m1-3 paraconid height
(0) low
(1) tall
9. m1-3 paraconid position
(0) distinctly separated from metaconid
(1) close to metaconid or fused with metaconid
10. Molar paracristid shape
(0) Buccal portion of paracristid directed lingually
(1) Buccal portion of paracristid directed mesially
11. m1-3 metaconid inflation
(0) Metaconid inflated
(1) Metaconid uninflated
(2) Metaconid reduced
12. m2 cristid obliqua orientation
(0) Nearly parallel to long axis of tooth
(1) Around 45 degrees to long axis of tooth
13. Position of m1-3 hypoconid
(0) Mesial to entoconid
(1) Aligned with entoconid
14. Position of m1-3 entoconids
(0) Mesial to distal margin of talonid
(1) At distal margin of talonid
15. m1-2 entocristid
(0) Present and ascends distal slope of metaconid to close talonid
(1) Present and terminates at base of metaconid
16. m3 relative size
(0) Unreduced, subequal to or larger than m2
(1) Reduced, noticeably smaller than m2
17. m3 talonid width
(0) Less than trigonid width
(1) Subequal to trigonid width
18. P4 metacone
(0) Absent
(1) Present
19. P4 protocone size
(0) Small or absent
(1) Subequal to paracone
20. P4 postprotocrista orientation
(0) Buccal
(1) Distal
21. M1-2 stylar shelf
(0) Reduced
(1) Absent
22. Orientation of M1-2 preparacrista (position of parastyle)
(0) Oriented mesiobuccally (parastyle buccal to para-, metacones)
(1) Approximately in line with the para- and metacones
23. M1-2 mesostyle
(0) Absent
(1) Present
24. M1-2 postprotocrista
(0) Present
(1) Absent
25. M1-2 postprotocingulum (Nannopithex fold)
(0) Absent
(1) Present
26. M1-2 hypocone size (Ordered)
(0) Reduced or absent
(1) Moderately developed
(2) Enlarged, subequal in size to protocone
27. M1-2 hypoloph
(0) Absent
(1) Present
28. Calcaneal tuber relative length
(0) short
(1) intermediate
(2) elongate
29. Calcaneal fibular facet
(0) Absent
(1) Present, relatively flat
(2) Present, describing an arc of ~180 degrees
30. Orientation of long axis of ectal facet
(0) Subparallel to long axis of calcaneum
(1) Approximately 45 degrees to long axis of calcaneum
(2) Nearly 90 degrees to long axis of calcaneum
31. Ectal facet curvature
(0) Gently, evenly convex
(1) Strongly convex
(2) Sharply divided into posterior and anterior surfaces
(3) Flat
32. Overlap between calcaneal ectal and sustentacular facets
(0) No overlap
(1) Partial overlap
(2) Nearly complete overlap
33. Distal extension of sustentacular facet
(0) Absent
(1) Present
34. Position of plantar tubercle
(0) At distal margin of calcaneum (level of cuboid facet)
(1) Shifted proximally
35. Peroneal tubercle position and size
(0) At distal margin of calcaneum, strongly reduced
(1) At distal margin of calcaneum, elongate
(2) Between distal margin and sustentaculum, elongate
(3) At level of sustentaculum, elongate
36. Curvature of cuboid facet
(0) Flat (no dorsoventral or mediolateral concavity)
(1) Concave dorsoventrally, flat mediolaterally
(2) Concave both dorsoventrally and mediolaterally
(3) Concave mediolaterally, flat dorsoventrally
37. Orientation of ML axis of cuboid facet relative to long axis of calcaneum (Ordered)
(0) ~90 degrees
(1) ~70-80 degrees
(2) less than ~70 degrees
38. Proportions of cuboid facet (Ordered)
(0) Facet much deeper (dorsoventral) than wide (mediolateral)
(1) Facet depth and width subequal
(2) Facet much wider (mediolateral) than deep (dorsoventral)
39. Medial trochlear ridge
(0) Absent (medial facet of trochlea faces dorsomedially)
(1) Present (medial facet faces medially)
40. Depth of trochlear groove
(0) Shallow
(1) Moderately deep (U-shaped)
(2) Deep (V-shaped)
41. Radius of curvature of lateral trochlear ridge
(0) Greater than medial ridge
(1) Subequal
42. Astragalar foramen
(0) Present
(1) Absent
43. Posterior extension of trochlear articular surface
(0) Absent, articular surface ends on dorsal surface
(1) Posterior extension of medial ridge
(2) Entire trochlea extends posteriorly
44. Cotylar fossa (Ordered)
(0) Weak to absent
(1) Prominent, dorsally oriented
(2) Prominent, medially oriented
45. Posteromedial projection of astragalar body
(0) Absent
(1) Present
46. Proximomedial plantar tuberosity
(0) Weak to absent
(1) Strong
47. Ectal facet width
(0) Transversely narrow
(1) Broad
48. Astragalar sustentacular facet shape and curvature
(0) Circular, axis of curvature subparallel to long axis of neck
(1) Elongate, axis of curvature perpendicular to long axis
(2) Elongate, wrapping around to med side of neck, axis subparallel
49. Orientation of long axis of navicular facet
(0) Essentially perpendicular to axis of trochlea
(1) Oblique to or parallel to axis of trochlea
50. Facet for medial tarsal/spring ligament
(0) Extends posteriorly along astragalar neck
(1) Does not extend far beyond level of navicular facet
51. Astragalar cuboid facet
(0) Present, nearly coplanar with navicular facet
(1) Absent
(2) Present, separated from navicular facet by a strong ridge
52. Navicular facet convexity
(0) Convex in one direction
(1) Flat
(2) Concave
Character-taxon Matrix:
Protungulatum
0010010100010010000000000000110110101210101000001000
Hyopsodus
0210110110010010101011000110111100021100101010000000
Apheliscus
110111100?001111100111100001210100011112012100001000
Haplomylus
1111101001010111010111001011222100011112012201001120
Paschatherium
0020110000111101010010000100011100131112012100001000
Mithrandir
0110110100011000000001000100110110101210101010001000
Chriacus
0100001000100010100-10000000111100220110101001000100
Ectocion
1210110010110110111010110201110111011112002000011000
Tetraclaenodon
12101101100101100110101001011-----1---1-00201----0-0
Claenodon
1101000110000110100011000101110110201110101000001000
Diacodexis
0101000110010000100011000001203201011012112000011022
“Hyracotherium”
1210110--1101010111011010201122011011012112000111022
Molinodus
1120-101100010110000110010102201001-1100-0101-00000-
Dissacus
00011100012011-1000011000001112201001012101010011022
Macroscelidea
111111-000011101010111011211222000010112112201001120
Hyracoidea
1211110001011110111110110210023001300112012100101111
5. Analyses performed and results.
All analyses were performed using the parsimony ratchet algorithm of NONAv2.0 (Goloboff, 1999) spawned by Winclada(BETA)v0.9.9 (Nixon, 1999b). See Nixon (1999a) for a discussion of the parsimony ratchet algorithm. Ten repetitions of each analysis were performed to ensure that the most parsimonious topology recovered by a given analysis represented the true shortest tree or trees.
5a. Menget al. (2003).
Meng et al.’s (2003) matrix, with the modifications above, was analysed two ways, one with all characters unordered and a second with those multistate characters listed above ordered. The ordered analysis produced sixteen most parsimonious trees of length 979 steps, with a consistency index of 39 and a retention index of 77. Haplomylus is identified as the sister taxon to extant Macroscelidea (represented by Rhynchocyon and Petrodromus), and Apheliscus forms the sister taxon to this clade. The full strict consensus tree is shown in Supplemental Figure 1. The unordered analysis produced twelve most parsimonious trees of length 926 steps, with a consistency index of 41 and a retention index of 76. In the unordered analysis, Haplomylus and Apheliscus form an unresolved trichotomy with extant Macroscelidea. Otherwise, the strict consensus of the unordered analysis differs from that of the ordered analysis only in the interrelationships of Rodentia, as traditionally defined, and is shown in Supplemental Figure 2.
Supplemental Figure 1. Strict consensus of 16 most parsimonious trees (L: 979; CI: 39; RI: 77) from ordered analysis of a modified version of Meng et al.’s (2003) matrix.
Supplemental Figure 2. Strict consensus of 12 most parsimonious trees (L: 926; CI: 41; RI: 76) from unordered analysis of a modified version of Meng et al.’s (2003) matrix.
5b. Ungulate analysis
The matrix was analysed with the five characters identified above ordered or unordered. When all characters were unordered, a single most parsimonious tree (Fig. 3) was recovered with a length of 180 steps, a consistency index of 38 and a retention index of 47 (all tree statistics were calculated by Winclada. A single tree with the same topology was recovered by the ordered analysis and a length of 181 steps, a consistency index of 38, and a retention index of 48.
5c. Asher et al. (2003).
Asher et al.’s (2003) matrix, with the modifications above, was analysed with two taxonomic samples, one with only Apheliscus, Haplomylus, and extant afrotheres included in the ingroup, and the second with other potential extinct afrothere relatives. In both analyses, all characters were treated as unordered. Analysis with the restricted taxonomic sample produced a single most parsimonious tree of length 261 steps, with a consistency index of 66 and a retention index of 82, which is shown in Supplemental Figure 3. The broader taxonomic sample also produced a single most parsimonious tree (Fig. 3c) of length 380 steps, with a consistency index of 45 and a retention index of 58. Where taxa overlap between the two analyses, the topology is the same.