Full Length Article
New Phylogenetic Analysis of the Family Elephantidae Based on Cranial-Dental Morphology†
Article first published online: 20 NOV 2009
Copyright © 2009 Wiley-Liss, Inc.
The Anatomical Record
Volume 293, Issue 1, pages 74–90, January 2010
How to Cite
Todd, N. E. (2010), New Phylogenetic Analysis of the Family Elephantidae Based on Cranial-Dental Morphology. Anat Rec, 293: 74–90. doi: 10.1002/ar.21010
This article is dedicated in the memory of Dr. Jeheskel Shoshani.
- Issue published online: 28 DEC 2009
- Article first published online: 20 NOV 2009
- Manuscript Accepted: 25 JUN 2009
- Manuscript Received: 16 AUG 2008
In 1973, Vincent Maglio published a seminal monograph on the evolution of the Elephantidae, in which he revised and condensed the 100+ species named by Henry Fairfield Osborn in 1931. Michel Beden further revised the African Elephantidae in 1979, but little systematic work has been done on the family since this publication. With addition of new specimens and species and revisions of chronology, a new analysis of the phylogeny and systematics of this family is warranted. A new, descriptive character dataset was generated from studies of modern elephants for use with fossil species. Parallel evolution in cranial and dental characters in all three lineages of elephants creates homoplastic noise in cladistic analysis, but new inferences about evolutionary relationships are possible. In this analysis, early Loxodonta and early African Mammuthus are virtually indistinguishable in dental morphology. The Elephas lineage is not monophyletic, and results from this analysis suggest multiple migration events out of Africa into Eurasia, and possibly back into Africa. New insight into the origin of the three lineages is also proposed, with Stegotetrabelodon leading to the Mammuthus lineage, and Primelephas as the ancestor of Loxodonta and Elephas. These new results suggest a much more complex picture of elephantid origins, evolution, and paleogeography. Anat Rec, 2010. © 2009 Wiley-Liss, Inc.
In 1973, Vincent Maglio published a seminal monograph on the evolution of the Elephantidae. In his phylogeny, three lineages of elephants, Loxodonta, Elephas, and Mammuthus, evolved from Primelephas in Africa, ∼6 ma. Not only did Maglio (1973) summarize the origin, evolution, and zoogeography of the entire family, he also consolidated the 100+ species proposed by Henry Fairfield Osborn in his two-volume Proboscidea into 25 valid species.
Beden (1979) expanded on Maglio's (1973) work, but focused on the African Elephantidae. He was responsible for the identification and description of material from East Lake Turkana, Kenya, the Omo Valley, Ethiopia, Laetoli, Tanzania, and Hadar, Ethiopia. These collections represent the bulk of the elephant material from Africa, and his collected works are a testimony to the effort and required to identify such a large amount of material in such a small amount of time.
In 1996, The Proboscidea, The Evolution and Palaeoecology of Elephants and Their Relatives (Shoshani and Tassy, 1996) was published as an update of proboscidean studies since Osborn's 1936 and 1942 Proboscidea volumes. Seventeen other species have been added to the list of taxa recognized by Maglio (1973). Some were considered to be junior synonyms by Maglio, but have now been re-evaluated and reinstated. Others are completely new species. Since 1996, three additional species have been described and a number of subspecies are now formally recognized. This brings the total number of elephantid species to 43, though not all of these are included in the cladistic analysis presented in this article (Table 1).
|Maglio (1970)||Shoshani and Tassy (1996)||Additional species/subspecies|
|Loxodonta adaurora||Loxodonta adaurora adauroraa|
|Loxodonta adaurora kararaea|
|Loxodonta atlantica||Loxodonta atlantica atlanticaa|
|Loxodonta atlantica zulu|
|Loxodonta africana||Loxodonta africana africana|
|Loxodonta africana cyclotis|
|Elephas recki Stages I-IV||Elephas recki brumptia|
|Elephas recki shungurensisa|
|Elephas recki atavusa|
|Elephas recki ileretensisa|
|Elephas recki reckia|
|Elephas namadicusb||Elephas (Paleoloxodon) antiquus|
|Elephas falconeric||Elephas (Paleoloxodon) falconeri|
|Elephas (Paleoloxodon) mnaidriensis|
|Elephas (Paleoloxodon) creutzburgi|
|Elephas (Paleoloxodon) naumanni|
|Elephas ? beyeri|
|Elephas (Paleoloxodon) creticus|
|Elephas (Paleoloxodon) cypriotes|
|Elephas maximus||Elephas maximus maximus|
|Elephas maximus indicus|
|Elephas maximus sumatranus|
|Mammuthus meridionalis Laiatico Stage, Montavarchi Stage, and Bacton Stage||Mammuthus meridionalis gromovi|
|Mammuthus meridionalis meridionalis|
|Mammuthus meridionalis vestinus|
|Mammuthus armeniacusd||Mammuthus trogontherii|
|Mammuthus colombie||Mammuthus? jeffersoni|
|# Species 25||40||43|
Both Maglio's (1973) and Beden's (1979) monographs were primarily concerned with specimen and species identification. Character evolution was determined by identifying an ancestor from the older fossils from which the evolution of more recent elephants could be deduced. The material recovered from late Miocene and early Pliocene African localities seemed to be the key to the origin of the family. Preoccupation with an ancestral morphotype, and the progressive development of characters from this morphotype to the extant elephants has resulted in a preoccupation with anagenetic trends. Subsequent to Maglio (1973), species and subspecies have been defined based on their position within these trends, rather than based on the morphological characteristics that they share or do not share with other species. As a result, traditional systematic methods have been unable to distinguish between homoplasies (which occur in parallel in all lineages of elephants), plesiomorphies, and synapomorphies.
Unfortunately, this emphasis on identification led to inconsistent use of terminology, character descriptions, and measurement methods. Many of the similarities in morphology among fossil elephants were based on plesiomorphic characters. Loxodonta has long been thought to represent the ancestral condition for the family, and yet many similarities have been observed between this genus and several late Pleistocene taxa in Europe (Todd, 1997). Unrelated species were grouped together based on superficial resemblances that were the result of parallel evolution or retention of ancestral characters (Maglio 1973). This is further complicated by the widespread parallel evolution that has occurred in all three lineages, Elephas, Mammuthus, and Loxodonta.
There have been few attempts to examine the relationships of species within the Elephantinae through character-based analysis. Tassy (1988, 1990), Tassy and Darlu (1986, 1987), and Kalb and Mebrate (1993) have examined the Order Proboscidea and/or the Suborder Elephantoidea using cladistic methods. Tassy (1988, 1990) and Tassy and Darlu (1986, 1987) have analyzed the relationships within the Proboscidea, particularly the relationships of tetralophodont-grade elephantoids to each other and to the Elephantidae. In his discussion of phylogeny and classification of the Proboscidea, Tassy (1990) reviews previous classifications and examines the relationships within the order using parsimony. His cladogram of the Proboscidea is resolved except for node 13 (the Elephantoidea defined as including the Mammutidae, Ambelodontidae, “gomphotheres,” Choerolophodon, Stegodontidae, and the Elephantidae). This is the crucial node for the origin of the elephantines (Stegodibelodon, Primelephas, Loxodonta, Mammuthus, and Elephas), and suggests that the origin of the Elephantidae is not as simple as may have been previously thought.
Previous research has been aimed at determining the probable ancestor of the family Elephantidae, the relationships of the sister groups Stegodontidae and Gomphotheriidae to the Elephantidae, and the status of the paraphyletic “gomphothere” group. As a result, many of the characters used in these analyses are too generalized to distinguish between species within the Elephantidae. The Elephantidae are included in these analyses at the generic level only, and no study of the species relationships within the family has been done.
Kalb and Mebrate (1993) analyzed characters at the generic level in African elephants, using proboscidean specimens from the Middle Awash, Ethiopia. This is the only recent study which focuses on dental characters of African Elephantidae using character-based analysis. They separate a “Loxodonta Group” from Loxodonta adaurora and discuss the relationships of both Loxodonta clades to a Primelephas-Mammuthus-Elephas clade. They grouped Primelephas with Mammuthus and Elephas based on synapomorphies (concave-concave enamel loops on lower molars, single posterior column). Twin posterior columns in Loxodonta adaurora and convex-concave enamel loops in the Loxodonta Group separate these two groups from the Primelephas-Mammuthus-Elephas clade (Kalb and Mebrate, 1993). The presence of a prominent posterior column which is completely isolated in unworn molars in these species reflects the ancestral “trefoil” pattern from a gomphothere ancestor, and thus explains the direction of the development of the median loop (Kalb and Mebrate, 1993).
Study of the morphology of the extant species (Loxodonta africana and Elephas maximus) provides the basis for a more rigorous analysis of the characters used to define fossil species and subspecies of Elephas, Loxodonta and potentially Mammuthus (Todd, 2009). The goal of this article is to present a cladistic analysis of the Elephantidae based on a new, descriptive cranial-dental character dataset, and to suggest a new phylogenetic scheme for the family.
MATERIALS AND METHODS
Species and Specimens
Because of the variability in specimens that have been assigned to fossil taxa in the Elephantidae, and the variability in descriptions of such taxa in the literature, characters were coded on the type specimens for each species and subspecies only. The type specimen data is proposed as the most accurate representation of the “true” morphology of each fossil and modern species. Most of the type specimens consist of molars only, but some do include the cranium and/or mandible. In a few cases, only the mandible and mandibular teeth exist.
A few of the type specimens do not exist anymore, or the specimen was inaccessible. The whereabouts of the type for Loxodonta africana (originally named Elephas africanus) is unknown. This specimen was examined using a drawing from Osborn (1942:1197), and another molar from the type locality (Cape Colony) located in the British Museum of Natural History. The only designated type for any of the subspecies of Loxodonta africana or Elephas maximus is “Congo,” the type specimen for Loxodonta africana cyclotis. This specimen was located in the American Museum of Natural History. An adult representative of each of the other extant subspecies was chosen as the specimen for use in coding character states.
The types for Loxodonta atlantica and Elephas iolensis are housed in the Musée Nationale d'Histoire Naturelle in Paris, but are listed without specimen numbers. These are figured in Osborn (1942), however, and the type localities for each are listed. The molars best matching both of these were examined during research in Paris. An excellent cast of the type for Mammuthus subplanifrons was examined in the British Museum. Only one specimen of Mammuthus meridionalis was studied, and this was used as the example for this species.
The lectotype for Loxodonta exoptata is from Laetoli, Tanzania, and is housed in the Humboldt Universität, Berlin. This specimen is figured by Maglio (1969:18, Plate I, Figs. 1 and 2), and the photograph was used in addition to a specimen from East Turkana, Kenya. The lectotype for Elephas recki (also Elephas recki recki) is from Olduvai Gorge, Tanzania and is located in the same museum. The specimen itself was not examined, but a cast reconstruction matching the figure from Osborn (1942) that was in the collections of the National Museum of Kenya was studied. The types for the subspecies Elephas recki brumpti, and Elephas recki shungurensis are from the Omo Valley, Ethiopia, and were not available for study. Figures of these from Beden (1979) and comparable specimens from other collections were used as the example for each. The type for Elephas recki atavus is on display in the Gallerie de Paléontologie in Paris. This is a very complete cranium with both upper third molars still in place. The mandible for Elephas recki atavus was not found, and so KNM-ER 5711 was used for mandibular characters. This specimen consists of a cranium and mandible that is virtually identical to the type specimen. The type specimen for Elephas ekorensis consists of upper right and left M3, and so cranial data was collected from KNM-EK 422, a partial cranium housed in the National Museums of Kenya. A list of the type specimens and specimens used to code for characters is included in Table 2.
|Stegotetrabelodon orbus||KNM-LT 354||Partial mandible with LLM2-and LLM3|
|Primelephas gomphotheroides||KNM-LT 351||ULM3, URM3, LM3, and fragmentary palate|
|Loxodonta adaurora and Loxodonta adaurora adaurora||KNM-KP 385||Almost complete adult skeleton, partial cranium with UM3, and complete mandible with LM3|
|Loxodonta adaurora kararae||KNM-ER 347||Partial cranium with UM3|
|Loxodonta atlantica||MNHN||Co-type, LRM2, 1 referred specimen-ULM3|
|Loxodonta africana||Location unknown, type locality probably Cape Colony, figured in Osborn (1942:1197)a||LRM2, used URM3 from Cape Colony locality, and NMNH 304615 for cranium and mandible characters|
|Loxodonta africana oxyotis||No type specimen||Used NMNH 304615|
|Loxodonta africana cyclotis||AMNH 90102 (?) “Congo”||Complete cranium and mandible M2 in wear and M3 forming|
|Loxodonta exoptata||IPUB Z.94.96, figured in Maglio (1969:18)a||LRM3, also used KNM-ER 3200 A-B|
|Elephas ekorensis||KNM-EK 424||URM3, ULM3, also used KNM-EK 422 (cranium)|
|Elephas recki and E. r. recki||IPUB XVII 1382b||LLM2 in mandible fragment, examined cast of this specimen|
|Elephas recki brumpti||Omo L1.33, figured in Beden (1979:387)a||Mandible fragment with LLM2 and LLM3|
|Elephas recki shungurensis||Omo 148.72.1, figured in Beden (1979:397)a||Maxilla fragment with ULMa and ULM3|
|Elephas recki atavus||MNHN 1933.9.300||Complete cranium with URM3 and ULM3|
|Elephas recki ileretensis||KNM-ER 1588||Maxilla fragment with URMb and URMa|
|Elephas antiquus||BM M.2006||LLM2|
|Elephas namadicus||BM M.3092||Partial cranium with frag. UM3|
|Elephas hysudricus||BM M.3109||Cranium|
|Elephas meletensis||BM M.44312||ULMa(?)|
|Elephas mnaidriensis||BM M.44304||LRM3|
|Elephas maximus maximus||No type specimen||Used ROM 01.2.8.1|
|Elephas maximus indicus||No type specimen||Used AMNH 30249|
|Elephas maximus sumatranus||No type specimen||Used NMNH 282837|
|Mammuthus subplanifrons||MMK 3020, cast in BMNHb||Type of “Archidiskodon subplanifrons”|
|Mammuthus meridionalis||IGF 1054a||Used MNHN 1948-1-126|
|Mammuthus armeniacus||BM 32250, BM 32252||ULM3 and URM3, also referred to “M. trogontherii”|
Data sets were analyzed using Hennig86. With the exception of one analysis, trees were obtained using the mhennig command with branch swapping (mh*; bb*) (trees were obtained for one dataset using ie*). For each data set, trees were obtained from ordered, successively weighted characters (cc; xsteps w) as well as unordered and successively weighted characters (cc-; xsteps w). Finally, multiple trees were condensed into a Nelsen consensus tree (n; tplot).
Seventy-seven multistate characters were examined on two species of extant elephants and 15 extinct species. This dataset includes a total of 77 multistate characters: 33 dental characters, 32 cranial characters, and 12 mandible characters defined based on the osteological analysis of the extant elephants, as well as the extinct species (Table 3) (Todd, 1997). Three general skeletal characters and seven phenotypic characters were also studied, but at the present time, these can only be used to identify subspecies of the two living elephants and so are not included in the cladistic analysis.
|1. Molar shape|
|0 = tapered at anterior end (ovate)|
|1 = parallel-sided|
|2 = widest in middle (elliptic)|
|3 = tapered at posterior end (ovate)|
|2. Molar curvature|
|0 = straight|
|1 = curved at posterior end|
|3. Molar shape|
|0 = height even at both ends|
|1 = greatest height at posterior end|
|*Generally characterizes upper and lower teeth|
|4. Greatest tooth width|
|0 = base of crown|
|1 = 1/4 up from base of crown|
|2 = 1/2 up from base of crown|
|3 = crown (posterior view)|
|5. Molar crown|
|0 = ends at alveolar border|
|1 = extends below alveolar border|
|0 = present|
|1 = absent|
|7. Occlusal surface|
|0 = even|
|1 = twisted|
|2 = sagging in middle|
|8. Inclination of plates to occlusal surface|
|0 = weak|
|1 = strong|
|9. Valleys between plates|
|0 = V-shaped|
|1 = U-shaped|
|10. Valley shape at base|
|0 = compressed, diverge at apex|
|1 = parallel|
|11. Cement filling valleys|
|0 = no|
|1 = yes|
|12. “S” curve to plates|
|0 = no|
|1 = yes|
|13. Lateral edges of plate|
|0 = low and rounded|
|1 = straight, angled in toward apex|
|2 = parallel-sided|
|3 = high and bowed out slightly|
|14. Molar roots|
|0 = strong or bifurcated|
|1 = absent or open|
|15. Apical digitations|
|0 = few (4 or less)|
|1 = many (greater than 4)|
|16. Appearance of complete enamel loops|
|0 = slow (within 6 worn plates)|
|1 = quick (within 3 worn plates)|
|17. Single column at posterior end|
|0 = present|
|1 = small plate|
|*May be variable|
|18. Anterior/Posterior columns|
|0 = strong anterior column|
|1 = strong posterior column|
|2 = strong anterior and posterior columns|
|3 = no anterior/posterior columns|
|19. Median cleft|
|0 = strong|
|1 = weak|
|2 = absent|
|20. Tusk shape|
|0 = straight|
|1 = curved or spiralled in front|
|2 = straight spiral (twisted)|
|21. Tusk cross-section|
|0 = rectangular or flattened|
|1 = oval or “bean” shaped|
|2 = round|
|22. Enamel height above cement|
|0 = Low|
|1 = High|
|*May be related to wear and amount of abrasion|
|23. Enamel figure shape|
|0 = parallel-sided|
|1 = true lozenge|
|2 = parallel-sided with median loop|
|3 = “pseudo-lozenge”|
|4 = “keyhole” shaped|
|5 = rounded loops|
|24. Median area|
|0 = loop|
|1 = fold|
|2 = absent or open|
|25. Lateral sides of enamel figure|
|0 = pinched|
|1 = rounded|
|2 = intermediate|
|3 = rectangular|
|26. Direction-lateral sides of enamel|
|0 = turn anterior|
|1 = turn posterior|
|2 = even|
|*May be variable|
|27. Symmetry of enamel figure|
|0 = symmetrical, in line with long axis of molar|
|1 = asymmetrical, offset from long axis of molar|
|28. Medial edges of enamel figures|
|0 = separated|
|1 = in contact|
|29. Enamel folding|
|0 = absent|
|1 = regular|
|2 = irregular|
|3 = undulating|
|4 = crinkled|
|30. Placement of folds|
|0 = median area only|
|1 = entire length of enamel figure|
|2 = absent|
|31. Amplitude of enamel folding|
|0 = absent|
|1 = high|
|2 = low|
|32. Spacing between enamel folds|
|0 = absent|
|1 = tight|
|2 = loose|
|33. Crenated versus smooth enamel|
|0 = Smooth|
|1 = Crenated|
|*May be taphonomic|
|34. Parietal/Occipital crest (=nuchal ridge)|
|0 = pronounced ridge|
|1 = ridge|
|2 = smooth|
|35. Shape of nares opening|
|0 = “dumbell” shaped|
|1 = turned down at lateral edges|
|2 = rounded and turned up at lateral edges|
|*May be related to sexual dimorphism (frontal view)|
|36. Borders of nares opening|
|0 = sharp and pronounced|
|1 = smooth and rounded|
|37. Center of nuchal ridge|
|0 = smooth and even|
|1 = heart-shaped|
|2 = concave|
|38. Position of orbits relative to tooth row|
|0 = anterior to tooth row|
|1 = even with beginning of toothrow|
|2 = posterior to beginning of tooth row|
|39. Slope of forehead and premaxillaries|
|0 = premaxillaries steeper than forehead|
|1 = in same plane|
|2 = forehead steeper than premaxillaries|
|40. Temporal line|
|0 = smooth|
|1 = line|
|2 = ridge|
|41. Parietal depression|
|0 = absent|
|1 = muscle marking|
|2 = furrow|
|0 = bulbous|
|1 = flat|
|43. Slope of occipitals from condyles|
|0 = anterior|
|1 = vertical|
|2 = posterior|
|44. Position of supraoccipital relative to squamosal|
|0 = directly superior|
|1 = lateral|
|2 = medial|
|45. Alveolar border of premaxillaries|
|0 = even|
|1 = slopes down laterally|
|0 = parallel|
|1 = flared|
|2 = straight-sided but diverging|
|*May be related to sexual dimorphism|
|47. Shape of occipital condyles|
|0 = round or square|
|1 = triangular, elongated triangle|
|2 = “bean” shaped|
|48. Condylar fossa|
|0 = flat and wide|
|1 = flat and narrow|
|2 = deep and wide|
|3 = deep and narrow|
|49. Condylar facet|
|0 = even with fossa|
|1 = slopes posteriorly from fossa|
|2 = slopes anteriorly from fossa|
|0 = plate-like, with pronounced edges|
|1 = smooth, completely fused|
|51. Basio-occipital and vomer|
|0 = meet|
|1 = separated|
|52. Position of exoccipital relative to condyles|
|0 = lateral|
|1 = anterior|
|0 = rounded|
|1 = flat|
|2 = concave|
|54. Tusk sheaths|
|0 = rotated anteriorly|
|1 = even|
|2 = rotated posteriorly|
|55. Ventral depression of palate|
|0 = no|
|1 = yes|
|56. Position of occipital condyles relative to tooth row|
|0 = posterior|
|1 = in line with end of tooth row|
|57. Opening of external nares|
|0 = above orbit|
|1 = even with orbit|
|2 = below orbit|
|58. Anterior portion of zygomatic|
|0 = projecting anteriorly|
|1 = receding|
|0 = Directed out and forward|
|1 = Directed out and downward|
|60. Height of occipital condyles|
|0 = low|
|1 = elevated|
|61. Frontal bones|
|0 = elongated|
|1 = shortened|
|62. Skull shape|
|0 = rounded|
|1 = compressed parallel to facial plane|
|63. Palate keel|
|0 = absent|
|1 = present|
|64. Nuchal fossa|
|0 = deep|
|1 = shallow|
|65. Occipital bosses|
|0 = in facial plane|
|1 = overhang forehead|
|66. Mandibular condyle shape|
|0 = long anterio-posteriorly|
|1 = oval|
|2 = rectangular|
|67. Condyle surface|
|0 = slopes medially|
|1 = even|
|68. Ascending rami|
|0 = diverging|
|1 = curve in|
|2 = parallel|
|69. Mandibular symphysis|
|0 = directed forward|
|1 = directed down|
|70. Length of mandibular symphysis|
|0 = long|
|1 = intermediate|
|2 = short|
|71. Shape of symphyseal trough|
|0 = horseshoe shaped|
|1 = U-shaped|
|2 = V-shaped|
|72. Mandibular corpus|
|0 = swollen|
|1 = gracile|
|73. Position of coronoid process relative to maximum length of corpus|
|0 = posterior|
|1 = half way|
|2 = anterior|
|74. Lateral side of ascending ramus|
|0 = concave|
|1 = flat|
|75. Angle of ascending ramus relative to corpus|
|0 = 90° angle|
|1 = acute angle|
|76. Height of ascending ramus relative to maximum length of corpus|
|0 = ramus height < corpus length|
|1 = ramus height = corpus length|
|2 = ramus height > corpus length|
|77. Lower incisors|
|0 = present|
|1 = germ cavity|
|2 = absent|
Once the character set was established, the literature was surveyed in detail for comparison of character descriptions. This was necessary to insure consistent terminology. Previous studies by Maglio (1973) and Beden (1979, 1983, 1987) provided the groundwork for interpreting differing descriptions, but these descriptions are often not used consistently across taxa. This has made it difficult to compare species and specimens, and is the principle reason for developing a new, well-defined character data set. Many of the new characters have a foundation in these previous studies, but all are new in terms of their descriptions and states.
Two examples of the comparison process between previous studies and the new analysis are the dental characters involving folding of the enamel and the shape of the enamel figure. Enamel folding is so variable in fossil elephants that it was necessary to separate this general character into five separate characters. The basic trend in dentition that has been proposed for all elephant lineages is increased shearing efficiency of the molars. This was accomplished in different ways in each lineage, but there are allometric and functional changes in several features which are intricately related. To increase plate spacing while maintaining numbers of plates, the enamel thickness had to decrease. To compensate for thinner (and less durable) enamel, the crown height increased. To compensate for thinner enamel, and maintain adequate surface area, enamel became increasingly folded. Although the degree and type of enamel folding varies among species, there are very definite characteristics which can be isolated as characters with several states each.
Enamel folding is so complicated in fossil elephants, that not only the pattern was coded but also four additional characters were created: placement of folds, amplitude of folds, spacing of enamel folds, and crenated versus smooth enamel. Every specimen was coded for these characters, and the results show a high degree of polymorphism within-species. Even when separated into five different multistate characters, there is still a wide range of variation.
A second example of comparisons with previous studies and the new analysis is the shape of the enamel figure. Again, the descriptions of this feature in the literature are varied and too comprehensive for it to be a single character. As described in the introduction, some of this variation is due to developmental plasticity, but is also related to shape changes in the molars, such as overall width and allometric changes in enamel and plate thickness. In this case, the original feature has been divided into six multistate characters: presence of anterior/posterior columns, general enamel figure shape, shape of median area, lateral edges of enamel figure, direction of lateral edges of enamel, and symmetry of median loop. As with the first example, these characters are still highly polymorphic.
The total data set includes 33 dental characters, 32 cranial characters, and 12 mandibular characters. The outgroup, Stegotetrabelodon orbus, and nine other species are included in this analysis. When characters are ordered, one tree was obtained with a Length = 176, CI = 65, and RI = 51 (Fig. 1). When the characters are unordered, 16 equally parsimonious trees are obtained with Length = 153, CI = 71, and RI = 53. A Nelson Consensus Tree is represented in Figure 2. The overall structure of the two final trees is the same, though the relationship of Eurasian Elephas to African Elephas is unresolved.
The dental data set contains 33 dental characters. The outgroup is Stegodon kaisensis and includes 18 other species. As this data set is based only on dentition, more species can be included. Stegodon kaisensis (Stegodontidae) is a member of the sister group to the Elephantidae, and provides a better outgroup than Stegotetrabelodon when all genera of the Elephantidae are included in the analysis. However, there is no cranial material for Stegodon kaisensis, so it cannot be used as an outgroup in the total character analysis.
Nineteen equally parsimonious trees were obtained with the characters ordered, Length = 151, CI = 43, RI = 58. One tree resulted with successive weighting of characters (Fig. 3). In an unordered analysis, two equally parsimonious trees were obtained with a Length = 120, CI = 50, and RI = 60. The only difference between the two trees is the position of S. orbus and P. gomphotheroides (Fig. 4).
Homoplasy is a persistent problem in this cladistic analysis. All three lineages are undergoing similar trends in character evolution, but at different rates and times. As a result, each character appears on the tree several times, resulting in low CI indices. In two clades from an analysis of subspecies and species (Todd, 2006), the dental similarities are clearly seen (Fig. 5,6). Some of these similarities have phylogenetic significance, (Elephas iolensis and Elephas hysudricus), while others illustrate the unique problems of parallel evolution, and the difficulty in sorting out homoplasy resulting from convergence (Loxodonta atlantica and Elephas recki recki). This creates a problem using parsimony, but it is possible to make conclusions about phylogenetic relationships if this complication due to parallel evolution is kept in consideration, and a new phylogeny is presented in Fig. 7.
Loxodonta is paraphyletic in both analyses of cranial and dental characters. Elephas is polyphyletic in the ordered analysis but paraphyletic in the unordered analysis due to the inclusion of Mammuthus primigenius with the Eurasian Elephas species. In the analysis of dental characters only, Loxodonta is paraphyletic in the ordered analysis but polyphyletic in the unordered analysis. Mammuthus is also paraphyletic in both analyses. There are many other species of Mammuthus that could be included in this analysis, but these are middle to late Pleistocene forms that are highly derived. The three species included here are the older, more ancestral species. Both Mammuthus meridionalis and Mammuthus armeniacus remain together as a clade, but do not group with Mammuthus subplanifrons. Based on the results of previous metric analysis of the family elephantidae, M. subplanifrons is inseparable from Loxodonta adaurora, an early loxodont species in Africa (Todd, 1997). Considering these results, and that M. subplanifrons consistently groups with L. adaurora as a basal member of the Elephantidae in the cladistic analysis, it is highly likely that these two species are the same species. This partially solves the homoplastic problem of Loxodonta, at least in the unordered analysis. There are certain features of the skull of L. adaurora which are similar to later mammoths including massive, flaring premaxillaries, twisted tusks (although straight), and a relatively flat frontal. This suggests that L. adaurora does not belong in the Loxodonta lineage, and may, combined with M. subplanifrons, represent the primitive mammoth condition.
The grouping of Stegotetrabelodon with L. adaurora supports Beden's (1979) phylogeny (Fig. 8). Both Maglio and Beden proposed Primelephas as the ancestor of Elephas and Mammuthus (and Maglio included Loxodonta as a descendant) and the position of Primelephas gomphotheroides as ancestral is supported in the cladistic analysis. However, the cladistic analysis, and previous metric analysis (Todd, 1997) suggests a closer relationship between Stegotetrabelodon and L. adaurora. Combined with the conclusion stated previously that M. subplanifrons and L. adaurora are the same species; this suggests that Stegotetrabelodon is ancestral to Loxodonta and Mammuthus. Primelephas is ancestral to Elephas only (Fig. 8). Both Maglio and Beden divided Elephas recki into smaller taxonomic units, Stages 1–4 by Maglio, Elephas recki brumpti, E. r. shungurensis, E. r. atavus, E. r. ileretensis, and E. r. recki by Beden. There is still much variation even within these groups, and E. recki as proposed is not a valid species (Todd, 2005). There are specimens that belong to Loxodonta and Mammuthus, as well as Elephas and the Paleoloxodon group that had been previously attributed to E. recki. Originally, a subgenus of Elephas, Paleoloxodon is resurrected in this new phylogeny, and includes specimens previously attributed to E. recki from 2.5 to 4 ma in Africa. Subspecies designations for E. recki are no longer considered valid (Fig. 7).
In Maglio's (1973) phylogeny, he includes one migration out of Africa for Elephas, that subsequently underwent an adaptive radiation in Eurasia (Fig. 9). Beden (1979) proposed two migrations of Elephas out of Africa; an earlier wave that evolved into Elephas (Elephas), and a second, younger wave that he designates Elephas (Paleoloxodon) due to similarities of these species to the African Elephas recki lineage. In this analysis, the Eurasian Elephas species, Elephas namadicus and Elephas antiquus, group together with the extant Elephas maximus and with Elephas meletensis consistently in both ordered and unordered cladograms. E. recki and Elephas iolensis (African species) group consistently with Elephas hysudricus (Eurasian) and Elephas mnaidriensis (Mediterranean dwarf). The cladistic analysis presented here supports Beden's phylogeny of two separate migrations out of Africa for Elephas, with potentially two separate dwarfing events in the Mediterranean groups, although the allometric issues involved with reduction in body size in elephants is potentially generating homoplastic noise and this needs further study.
Maglio (1973) and others firmly place E. hysudricus as the ancestor to E. maximus. In this analysis, E. maximus does not form a clade with E. hysudricus and E. iolensis. However, previous metric data supports the relationship, and it is retained here for now (Todd, 1997). The wide range of metric variation in dentition in the African Elephas group, as well as the Eurasian species complicates the situation, and much further work is needed. The retention of a lozenge-like shape to the enamel figure in E. recki, E. antiquus, and E. namadicus, as well as with the dwarfed Mediterranean species, suggests an evolutionary relationship, as do various cranial similarities, such as large parietal bosses that overhang the forehead. There are recent data supporting the inclusion of some of the Mediterranean dwarfs in Mammuthus (Poulakakis et al., 2002), and this is an interesting possibility.
Elephas hysudricus, Elephas planifrons, Elephas maximus, and other Eurasian species have parallel-sided enamel figures, much thinner enamel and more lamellae. Thus, there may be support here for two separate genera within the Elephas group, an Elephas group that migrates out of Africa ∼3.7 ma, and potentially back to Africa with Elephas iolensis, and a Paleoloxodon group that leaves Africa ∼2.5 ma (Fig. 7).
This cladistic study relies heavily on dentition, which appears to be highly variable in fossil species and lineages, and again, more analysis is needed on cranial characters which suggest specific evolutionary relationships between African and Eurasian species. In addition, examination of postcranial elements potentially adds further data for comparisons. This is the first cladistic analysis to focus specifically on the Elephantidae as a whole, and the first new phylogenetic proposal in many years. Much comparative study remains to be done, as well as analysis of metric variation in fossil and modern elephants in order to further elucidate the complicated evolutionary relationships within this family.
This study would not have been possible without the permission of the Directors and Curators at the following museums: American Museum of Natural History, National Museum of Natural History, National Museums of Kenya, Tanzanian Museum, Musée Nationale d'Histoire Naturelle, British Museum of Natural History, and the Royal Ontario Museum. In addition, the author thanks many who helped with comments and information along the way, particularly P. Tassy, W. Sanders, M. Leakey, T. Harrison, J. Kalb, J. Kingston, A. Lister, L. Agenbroad, E. Sargis, T. Plummer, R. Potts, K. Behresmeyer, A. Brooks and D. Lipscomb.
- 1991. Fejej: a new paleoanthropological research area in Ethiopia. J Hum Evol 21: 137–143. , , , , , .
- 1981. The Miocene fossil beds of Maboko Island, Kenya: geology, age, taphonomy, and paleoecology. J Hum Evol 10: 35–48. , , , , .
- 2001. New data on the diversity of elephants Mammalia, Proboscidea in the early and early Middle Pleistocene of France. Proceedings of the World of Elephants, International Congress, Rome, 2001, p 81–84. .
- 1979. Les éléphants Elephas et Loxodonta d'Afrique orientale: systématique, phylogénie, intéret biochronologique, Thèse de Doctorat d'Etat, Poitiers. .
- 1981. Les proboscidiens des grands gisements à hominidés Plio-Pléistocènes d'Afrique Orientale. In: L'Environment des Hominidés au Plio-Pléistocène, Colloque International Organisé par la Fondation Singer-Polignac, New York. .
- 1987. Les Faunes Plio-Pléistocènes de la Vallée de l'Omo Éthiopie, Tome 2: Les Eléphantidés, Cahiers de Paléontologie, Paris. .
- 2002. Geology and geochronology of the middle Miocene Kipsaramon site complex, Muruyur Beds, Tugen Hills, Kenya. J Hum Evol 42: 11–38. , , , , .
- 1995. The Pleistocene locality of Kanjera, Western Kenya: stratigraphy, chronology and paleoenvironments. J Hum Evol 29: 247–274. , , , , , .
- 1997. Late Pliocene faunal turnover in the Turkana Basin, Kenya and Ethiopia. Science 278: 1589–1594. , , , .
- 1992. A new evaluation of the significance of the Late Neogene Lusso Beds, Upper Semliki Valley, Zaire. J Hum Evol 22: 505–517. , , , , , , , , , , , , , , , .
- 1995. The potential of the Turkana basin for paleoclimatic reconstruction in East Africa. In: VrbaES, Denton, GH, PartridgeTC, BurckleLH, editors. Paleoclimate and evolution, with emphasis on human origins. New Haven: Yale University Press. p 319–330. .
- 1996. The Pleistocene Dwarf Mammoths of Mediterranean Islands. In: ShoshaniJ, TassyP, editors. The proboscidea, evolution and palaeoecology of elephants and their relatives. United Kingdom: Oxford University Press. p 234–239. , , , .
- 1999. Browsing and grazing in elephants: the isotope record of modern and fossil proboscideans. Oecologia 120: 364–374. , , .
- 1993. Fossil proboscidean remains from Bolt's Farm and other Transvaal cave breccias. Palaeon Afr 30: 25–34. .
- 1970. Pleistocene Mammals from the Kaiso Formation and other related deposits in Uganda. In: LeakeyLSB, SavageRJG, editors. Fossil vertebrates of Africa, Vol 2. New York: Academic Press. , .
- 1972. Un nouveau proboscidean de Pliocène du Tchad, Stegodibelodon schneideri nov. gen. nov. sp., et le phylum des Stegotetrabelodontinae. CR Acad Sci Paris 274: 2962–2965. .
- 1978. Proboscidea. In: MaglioVJ, Cooke, HBS, editors. Evolution of East African Mammals. Massachusetts: Harvard University Press. p 333–367. , , , .
- 1995. A new species of Numidotherium Mammalia: proboscidea from the Eocene of Libya and the early phylogeny of the Proboscidea. J Vert Paleo 153: 650–671. .
- 2003. Mitochondrial cytochrome b of the Lyakhov mammoth Proboscidea, Mammalia: new data and phylogenetic analyses of Elephantidae. Mol Phylogenet Evol 26: 421–434. , , .
- 1988. Vertébrés du Pléistocène de Djibouti. Bull Soc Geol Fr 8: 323–334. , , , .
- 2002. 40Ar39Ar dating of Chemeron Formation strata encompassing the site of hominid KNM-BC 1, Tugen Hills, Kenya. J Hum Evol 421–422: 141–152. , .
- 1987. Remains of Proboscidea from the early Pliocene of Sahabi, Libya. In: BoazNT, El-ArnautiA, GaziryAW, de HeinzelinJ, BoazDD, editors. Neogene paleontology and geology of sahabi. New York: Alan R. Liss. p 183–203. .
- 1996. A Palaeocene proboscidean from Morocco. Nature 383: 68–70. , , .
- 1998. Phosphatherium escuillei du. Thanétien du Bassindes Ouled Abdoun Maroc, plus ancien proboscidien Mammalia d'Afrique. Géobios 30: 247–269. , , , .
- 2002. A new large mammal from the Ypresian of Morocco: evidence of surprising diversity of early proboscideans. Acta Palaeontologica Polonica 47: 493–506. , , , , , .
- 1997. Paleontology and biochronology of fossil localities in the Manonga Valley, Tanzania. In: HarrisonT, editor. Neogene Paleontology of the Manonga Valley, Tanzania. New York: Plenum Press. p 361–393. , .
- 1995. Faunal and environmental change in the Neogene of East Africa: evidence from Tugen Hills sequence, Baringo District, Kenya. In: VrbaES, DentonGH, PartridgeTC, BurckleLH, editors. Paleoclimate and evolution with emphasis on human origins. New Haven: Yale University Press. p 178–196. .
- 2002. New cercopithecoids and a hominoid from 12.5 ma in the Tugen Hills succession, Kenya. J Hum Evol 42: 75–93. , , , .
- 1999. The origin of grass-dominated ecosystems. Ann Missouri Botanical Garden 86: 590–564. , , .
- 1982. Vertebrate faunas from the Awash Group, Middle Awash Valley, Afar, Ethiopia. J Vert Paleo 2: 237–258. , , , , , , , , , .
- 1993. Fossil Elephantoids from the Hominid-Bearing Awash Group, Middle Awash Valley, Afar Depression, Ethiopia. Trans Am Phil Soc 83: xv-114. , .
- 2003. Laetoli paleoecology reconsidered: the isotopic evidence. Abstracts Paleoanthropology Society 2003 Meeting. Available at: http://www.paleoanthro.org/abst2003.htm. , .
- 2002. Stratigraphy, age and environments of the late Miocene Mpesida Beds, Tugen Hills, Kenya. J Hum Evol 42: 95–116. , , , .
- 1973–1974. Environment and subsistence of prehistoric man in the Southern Cape Province, South Africa. World Archaeol 5: 249–284. .
- 2005. Paléontologie systématique Paléontologie desVertébrés Réhabilitation de Mammuthus intermedius Jourdan, 1861, un mammouth Mammalia, Elephantidae du Pléistocène moyen récent d'Europe. CR Palevol 4: 235–242. , .
- 1996. Lothagam: a record of faunal change in the late Miocene of East Africa. J Vert Paleo 16: 556–570. , , , , , , , , , .
- 1995. New four-million-year-old hominid species from Kanapoi and Allia Bay, Kenya. Nature 376: 565–571. , , , .
- 1996. Evolution and taxonomy of the Eurasian Mammoths. In: ShoshaniJ, TassyP, editors. The proboscidea, evolution and palaeoecology of elephants and their relatives. United Kingdom: Oxford University Press. p 201–213. .
- 2005. The pattern and process of mammoth evolution in Eurasia. Quaternary Int 126–128: 49–64. , , , .
- 1970a. Early Elephantidae of Africa and a tentative correlation of African Plio-Pleistocene Deposits. Nature 225: 328–332. .
- 1970b. Four new species of Elephantidae from the Plio-Pleistocene of northwestern Kenya. Breviora 341: 1–43. .
- 1972. Evolution of Mastication in the Elephantidae. Evolution 26: 638–658. .
- 1973. Origin and evolution of the elephantidae. Trans Am Philos Soc 633: 1–149. .
- 1994. Expansion and emergence of C4 plants. Nature 371: 112–113. , , .
- 2002. Upper Siwalik mammalian faunas of India and associated events. J Asian Earth Sci 21: 47–58. .
- 1931. Palaeoloxodon antiquus italicus sp. nov. final stage in the Elephas antiquus phylum. Am Museum Novitates 460: 1–24. .
- 1936. The proboscidea: a Monograph of the discovery, evolution, migration and extinction of the mastodonts and elephants of the world, Vol. 1. Moeritherioidea, Deinotherioidea, Mastodontoidea. .
- 1942. The proboscidea: a Monograph of the discovery, evolution, migration and extinction of the mastodonts and elephants of the world, Vol. 2. Stegodontoidea, Elephantoidea. .
- 1988. Geology and fauna of the middle Miocene hominoid site at Muruyur, Baringo District, Kenya. J Hum Evol 3: 381–390. .
- 1989. Excavations and new findings at Kanjera, Kenya. J Hum Evol 18: 269–276. , .
- 2002. Origin and taxonomy of the fossil elephants of the island of Crete, Greece: problems and perspectives. Palaeogeograph Palaeoclimatol Palaeoecol 186: 163–183. , , , .
- 1990. Un nouveau gisement paléontologique du Plio-Pléistocène du Maroc: Ahl al Oughlam ancienne carrière Deprez. CR Acad Sci Paris sér II 310: 315–320. , , , .
- 2006. Complete mitochondrial genome and phylogeny of Pleistocene mammoth Mammuthus primigenius. PLoS Biol 43: e73. , , , , , , .
- 1990. Fossil proboscidea from the Pliocene Lusso Beds of the Western Rift, Zaire. Virginia Mus Nat Hist Mem 1: 171–187. .
- 2002. New proboscideans from the early Miocene of Wadi Moghara, Egypt. J Vert Paleo 222: 388–404. , .
- 1981. The Fort Ternan hominoid site, Kenya: geology, age, and paleontology. J Hum Evol 10: 49–72. , , , , .
- 1996. Para- or monophyly of the gomphotheres and their position within Proboscidea. In: ShoshaniJ, TassyP, editors. The proboscidea: evolution and palaeoecology of elephants and their relatives. New York: Oxford University Press. p 149–177. .
- ShoshaniJ, TassyP, editors. 1996. The proboscidea, evolution and palaeoecology of elephants and their relatives. United Kingdom: Oxford University Press.
- 2005. Advances in proboscidean taxonomy & classification, anatomy & physiology, and ecology & behavior. Quaternary Int 126–128: 5–20. , .
- 1976. The Lothagam 1 fauna: its phylogenetic, ecological and biogeographic significance. In: CoppensY, HowellFC, IsaacGL, LeakeyREF, editors. Earliest man and environments in the Lake Rudolf basin. Chicago: University of Chicago Press. p 361–369. .
- 1986. Nouveaux Elephantoidea Mammalia dans le Miocène du Kenya. Paris, France: Cahiers de Paléontologie. .
- 2003. In: LeakeyM, HarrisJM, editors. Lothagam: the dawn of humanity in Eastern Africa. New York: Columbia University Press. .
- 2001. The timing of early Elephantinae differentiation: the palaeontological record, with a short comment on molecular data. Proceedings of the World of Elephants, International Congress, Rome, p 695–687. , .
- 2003. Proboscidea and Tubulidentata. In: HarrisJ, LeakeyMEG, editors. Lothagam: the dawn of humanity in Eastern Africa. New York: Columbia University Press. , , .
- 1997. Comparison of variation in cranial and dental morphology of Elephas recki and the extant elephants: implications for phylogeny and macroevolution. PhD Dissertation, University Microfilms, George Washington University. .
- 1999. Fossil elephant diversity and its relevance to hominid evolution. Am J Phys Anthropol 28 ( Suppl): 265. .
- 2005. Reanalysis of African Elephas recki: implications for time, space and taxonomy. Quaternary Int 126–128: 65–72. .
- 2006. Proboscidean Diversity in the African Cenozoic. J Mamm Evol 131: 1–10. .
- 2009. Qualitative comparison of the cranio-dental osteology of the extant elephants, Elephas maximus (Asian elephant) and Loxodonta africana (African elephant). Anatomical Record (In Press). .
- 1996. Origin and radiation of the Elephantidae. In: ShoshaniJ, TassyP, editors. The proboscidea, evolution and palaeoecology of elephants and their relatives. United Kingdom: Oxford University Press. p 193–202. , .
- 1995. Paleoclimate and evolution, with emphasis on human origins. New Haven: Yale University Press. , , , .
- 1984. Hadar biostratigraphy and hominid evolution. J Vert Paleo 4: 575–581. , , .
- 1994. Ecological and temporal placement of early Pliocene hominids at Aramis. Ethiopia Nat 371: 330–333. , , , , , , .