table of contents
September 2001
[back] [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [next]

The Emergence of Whales, Chp. 11

Synopsis of Chapter 11: Implications of Vertebral Morphology for Locomotor Evolution in Early Cetacea

Emily A Buchholtz, Dept. of Biological Sciences, Wellesley College, Wellesley, MA

1. Introduction

3rd sentence: “The transition from a terrestrial habitat and paraxial, oscillatory limb locomotion to an aquatic habitat and axial undulatory locomotion was accompanied by marked changes (!!) in postcranial anatomy.”

Last sentence of first paragraph: “Although almost invariably fragmentary, early cetacean postcranian skeletons are surprisingly informative, and can complement morphological indicators of diet, sensory specializations, and skull reorganizations (!!) to fill in some of the gaps in our understanding of the dramatic transition (!!) of early cetaceans as they moved from land to water.”

First sentence, 2nd paragraph: “Cetaceans possess a large suite of derived traits associated with a marine lifestyle.”

Compare to mesonychids. Apparently Pachyaena ossifraga (written up in 1995) is the best choice out there — Gingerich’s article on Rodhocetus in Nature has a diagram of the skeleton).

Vertebral column census for P. ossifraga: 7 cervical, 12 thoracic, 7 lumbar, 3 sacral, and 15+ caudal vertebrae. Anterior thoracics indicate support for a large head. More detail follows, including indications of considerable range of tail movement.

Cetaceans are a lot different! Vertebral column census for Delphinus delphis (a dolphin, of course): 7 cervical, 13 thoracic, 22 lumbar, and 31 caudal vertebrae. (And creationists say that evolution doesn’t add anything new! How about 15 new lumbar vertebrae?) Cervical vertebrae are very short (P. ossifraga has long cervicals). More detail. Most motion occurs at the anterior (anal) and posterior (fluke) caudal nodes.

Primary database for the paper: fossil cetacean vertebrae. Will also compare to living aquatic tetrapods for motion similarities and differences.

2. Vertebral Variation and Locomotion in Aquatic Tetrapods

2.1 Functional Units of the Column

Changes in structure indicate changes in function. % change in a dimension (such as centrum length, width or height) is used to find the transition points. Example is Lutra canadensis (otter). The largest relative changes in Lutra’s vertebral dimensions occur at or immediately adjacent to traditional vertebral series boundaries (7/8, 21/22, 26/27, and 31/32 — diagram is good here).

Comparison is to the beluga whale. It’s hard to compare. (Also has a diagram). Maximum variation occurs at the caudal/thoracic boundary, midway in the thoracics, and at two points in the caudal series. (Divisions are thus: neck, chest, torso, peduncle, and fluke). Requires alternative terms for functional units.

2.1.1. Recognition of the functional torso

Living whales: posterior thoracics, lumbar, sacral, and anterior caudal vertebrae are similar and comprise the torso functional unit. Because sacrals are hard to recognize, it’s hard to identify the first true caudal. Typically considered to be the first vertebra with a hemal arch on the anterior.

2.1.2. Recognition of the peduncle and fluke

3 functional groups in the caudal vertebrae: anterior “lumbarized” region (functional torso); middle laterally compressed peduncle (makes the tail work efficiently); propulsive fluke.

Comparison of dolphin to manatee, because manatee has less well-defined peduncle and fluke. They’re different.

Cetacean evolution is “more accurately interpreted as a relocation of areas of dimensional discontinuity” from the traditional vertebrate locations to areas within series.

2.2 Changes in Cetacean Vertebral Counts

Cetaceans have more vertebra than terrestrial ancestors; the increase is distributed unequally in the different functional portions of the column.

Interestingly, Slijper (1936), considering only Protocetus, and both terrestrial mammals and cetaceans that he considered “primitive”, predicted that an “ancestral whale” would have 7 cervical, 17-19 noncervical presacrals (12-13 thoracics, 5-6 lumbars) and 2 sacrals. Quote: “Sixty years and many fossil discoveries later, these predictions have been largely upheld.”

[The editor of this series now takes the opportunity to shout impolitely: WHO SAYS THAT EVOLUTION DOESN’T MAKE PREDICTIONS??? Thank you. Now back to our show.]

She notes possible exceptions (allocation of thoracics and lumbars in the presacral region, number of sacrals). Some archaeocetes have 1 or 2 more thoracics and sacrals. Cervical number is “stable” at 7.

Elongation is a fairly common evolutionary trend, by the way. It’s important to cetaceans to provide muscle attachment area. PAY ATTENTION NOW: “Increase in the vertebral count of the torso may be used as a rough measure of propulsive power, and thus of swimming ability, in archaeocetes.”

[Question for the peanut gallery: is it likely that improved swimming ability would be favored by natural selection in a population that is adapting to an aquatic habitat?]

“Reduction in the number of sacrals, loss of fusion between individual sacral vertebrae, and loss of articulation of sacral vertebrae with the pelvis [Georgiacetus] undoubtedly all occurred during early whale evolution.”

Tail with non-propulsive function became an elongate, robust organ with propulsive function. [WHO says that evolution doesn’t add new functionality?]

2.3 Relative Centrum Length is an Indicator of Column Flexibility

If only the centrum length increases, the total column length increases and so does the possible displacement of the column.

2.3.1. Shortened Cervical Vertebrae Reduce Neck Flexibility

Reasons: enhanced hydrodynamic efficiency, front-end stabilization (good when the propulsion is from the back). Shortening in cetaceans is entirely due to centrum length reduction.

Mammals with terrestrial or transitional aquatic habitats have greater cervical centrum length than height. Mammals with obligate aquatic habitat (includes Sirenians and Phocids) have smaller cervical centrum length than height. Thus, aquatic locomotion preceded shortening of the cervical vertebrae.

2.3.2 Location of the Undulatory Portion of the Torso

Discussion of how the spine works in quadrupeds. This is cool: “The distinctive peak in centrum length between the limb girdles corresponds to the portion of the column undergoing the undulatory wave.” (The spine undulates during locomotion, see?)

Swimming lizards have a different pattern of lateral waves. Their spines have “plateaus” of long, nearly uniform centra. (There is a C. Gans, 1975, reference. No idea if this is a relative of Emperor Gans.)

Compares to motions in otariids (sea lions, fur seals), phocids (earless seals — never thought about that “earless” bit before: a transitional state?) and sea otters. Each have different patterns of centrum length peaks.

Sirenians have a single broad peak in centrum length, in the entire postthoracic column. Dugong is a bit different (centra are shorter), and the dugong swims faster than the manatee.

Archaeocete and living whales also have a single peak in pre-fluke centrum length (= undulatory wave with variable amplitude). BUT there is marked variation in the location and anteroposterior extent of the peak along the column. Archaeocetes have a trend toward posterior translocation of the peak.

3. Vertebral Patterns and Locomotion in Archaeocetes

We’re going to cover a lot here, so I’ll try to keep it succinct.

3.1 Ambulocetus natans

Skeleton has only a few fragmentary vertebrae. Count therefore unknown; more thoracic than lumbar, sacrals articulating with robust pelvis. Cervical centrum length relatively long. 1 measurable lumbar centrum is broader and markedly longer than high. Suggestion of “foot-swimming” by Thewissen (discussed by Fish in chapter 10) are supported by the existing vertebrae. Likely only a marginally aquatic lifestyle; less aquatic and more terrestrial profiency than extant phocids.

3.2 Remingtoncetus cf. R. harudiensis

Known from several localities. Note that several vertebrae had been originally described for Indocetus ramani. Predicted vertebral count: 7 cervicals, 14 thoracics, 5 lumbars, 4 sacrals, ? caudals. Cervicals longer than high, lumbar have broad centra slightly longer than thoracics (= sacrals). Also has a robust hind limb. Indications are for land locomotion; cranial air sinuses indicate pressure compensation ability, torso was likely not very undulatory. Locomotion: lumbar flexion with foot propulsion, possible small help from tail. “Semiaquatic”.

3.3 Rodhocetus kasrani

A nearly complete spinal column was found. Cervicals shorter than high. Thoracics (13) outnumber lumbars (6?), are 1.5 times broader than high, no increase in length through series Lumbars are the most robust vertebrae. “Sacralized” vertebrae are unfused, but pelvic articulation is still present. Centrum height/width are discontinous across the lumbar/sacral transition, centrum length increases to sacral peak.

So (gotta love this): “Rodhocetus presents vertebral characteristics of both terrestrial and aquatic animals.” (Disbelievers should read that again.) Centrum width/height discontinuities are like quadrupeds, allow for terrestrial ability (and pelvic articulation). Increase in lumbar centrum length (peaking in sacrals) suggests an undulatory torso with several series included. Supported by shortened cervicals (and the reduced femur). Probably had a sleek tail, most propulsion from the body. Can’t tell about a fluke, but possible, similar to manatee.

3.4 Protocetus atavus

7 cervicals, at least 12 thoracics, only one preserved sacral. 7 known lumbars, uncertain as to total number. Vertebrae centrum dimensions show “marked similarity” to Rodhocetus. Length of postaxis cervicals is nearly identical, implies aquatic lifestyle. Some difference in lumbar and function. Indications are that Protocetus was more fully aquatic than Rodhocetus, due to longer torso (more lumbars) and elongation of posterior thoracics.

3.5 Georgiacetus vogtlensis

We’ve heard a lot about this guy already. Cervicals are shortened. 4 sacrals are not fused. Centrum width “modestly” exceeds centrum height in the anterior thoracics, “markedly” exceeds it in posterior thoracics, anterior lumbars. Since it has a lack of dimensional discontinuities from posterior thoracic to anterior caudal vertebrae, it had a long undulatory torso with caudal peak length. So it was definitely an axial undulator, NO terrestrial competence. “Sleeker” than Rodhocetus or Protocetus. Probably had a caudal peduncle with a propulsive fluke. A darned good swimmer. So the novel about Georgiacetus might be entitled “Last of the Protocetids” (my apologies to James Fenimore Cooper).

3.6 Zygorhiza kochii

An upper Eocene dorudontine basilosaurid, remember. Multiple partial skeletons available. 7 cervical, 15 thoracic, 15 lumbar (including sacral), 21 caudal vertebrae. Markedly foreshortened cervicals, increase gradually down the spine, peak in midtail, i.e. long functional torso. No fused sacrals (though they’re still identified), no features for pelvic articulation. Very similar discussion of aquatic ability to Georgiacetus, with a definite caudal peduncle and fluke, sleek body, large relative length of lumbars. Catch this last lovely sentence: “Of all of the specimens examined in this study, this animal has a vertebral profile, and by implication a locomotor mode, most like that of living nondelphinid odontocete cetaceans.” Remember, upper Eocene. Very similar to Dorudon atrox. I.e. these are in the main line of descent to extant whales.

3.7 Basilosaurus cetoides

Late Eocene; a whole lot of them. Cervical vertebrae are extensively foreshortened. 15 thoracic vertebrae, 7 posterior are elongated, ribs articulate only with corresponding vertebrae, so they are functional lumbars. 13 lumbar and 2 sacral (Kellogg) are dimensionally similar to posterior lumbars and anterior caudals, so that makes 15. First 6 caudals are also lumbar in dimension, part of the functional torso. Posterior thoracics, lumbars and anterior caudals are extensively and uniformly elongated. (Visit the Smithsonian.) Midcaudal vertebrae show typical peduncle pattern of modern whales.

As we know, B. cetoides was fully marine. Short cervicals, no dimensional discontinuity at sacrals, no sacral/pelvic articulation indicate no terrestrial ability. The pattern of centrum length is different than other archaeocetes, with a plateau of dimensionally uniform elongate vertebrae. Indicates an anguilliform swimming pattern (like limbless lizards) with dorsoventral orientation. Means maneuverability triumped over speed. Small peduncle and fluke means that the undulation of the torso provided most of the power.


(I’m beat, by the way). So I’m going to skip most of the supporting text to the morphological trend points.

1. Reduction in number of sacral vertebrae (due to lack of fusion enhancing axial undulation.

2. Loss of sacral/pelvic articulation. Retention of one vertebrae with pelvic articulation (Rodhocetus) suggests that intrasacral fusion was lost before pelvic articulation. Undulatory motion possible in animals with some terrestrial ability.

3. Reduction in relative length of cervical vertebrae. Improves hydrodynamics.

4. Elongation of the torso. Increases undulatory range for more power.

5. Posterior translocation of maximum dorsoventral displacement point

6. Modification of tail to add peduncle and fluke. Note primarily that Rodhocetus and Protocetus are considerably different than Georgiacetus, Zygorhiza, and Basilosaurus.

Functional changes in swimming styles can be superimposed on the provisional archaeocete phylogeny. Two trends: expansion of the column representing the undulatory torso, and posterior movement of the undulatory wave peak with change in propulsive surface.

Second “suite” of changes are in body proportions, tending toward (as you might guess) sleekness and streamlining (tapering to the tail).

I truly do wish I could provide figure 12 with caption, which summarizes this quite well.


She does this twice. I hope it’s clear that she’s described a likely sequence of locomotor styles “in the movement from land to water habitats”. She summarizes the five stages (it’s actually four, Basilosaurus went off on a tangent). Here are the “titles” to the stages and the representatives:

1. Terrestrial quadrupeds. Pachyaena.

2. Lumbar undulation with limb propulsion. Ambulocetus and Remingtoncetus.

3. Lumbo-sacro-caudal undulation/variable amplitude (sacral maximum). Rodhocetus, Protocetus. (Efficient aquatic and very limited terrestrial locomotion)

4. Lumbo-sacro-caudal undulation/variable amplitude (caudal maximum). Georgiacetus, Zygorhiza.

5. Lumbo-sacro-caudal undulation (anguilliform). Basilosaurus.

That’s was a lot to chew on. Is this next chapter more digestible? Get out your napkins for

“Structural Adaptations of Early Archaeocete Long Bones”

which says a whole lot about the adaptations necessary for aquatic existence.

[back] [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [next]

This page was last updated September 1, 2001.
It was reformatted and moved August 6, 2007
Copyright © 2001 by James Acker

table of contents
September 2001