Horse Splint Bones

"The distal end is very variable in size and may be a mere point; in other cases, especially in large draft horses, it may present a prolongation which is regarded as the vestige of the digital skeleton." — Septimus Sisson, The Anatomy of the Domestic Animals, 1914¹

"The entire MTIV can be removed with good results, but removal of more than 50% of MCII can lead to carpometacarpal instability and lameness because MCII bears a substantial axial load." — Equine Surgery, third edition, 2006²

Summary

The parent article, Vestigial Structures, separates the dead claim that vestiges are useless from the live claim that they are reduced copies of something larger in an ancestor, and gives the live claim three steps: homology, reduction, and an ancestral distribution on a tree built from other characters.  This article runs that test on the two thin bones down the back of every horse's cannon.  The fossil series that supplies the third step belongs to the sibling article The Horse Series; this one owns the living horse.

• What they are. Septimus Sisson's veterinary anatomy of 1914 counts three metacarpal bones in the horse, of which "only one, the third or large metacarpal bone, is fully developed and carries a digit; the other two, the second and fourth, are much reduced, and are commonly called the small metacarpal or 'splint' bones."¹ Each is a tapering three-sided rod bound to the cannon by an interosseous ligament, with an enlarged head above that meets the carpal bones and a free button below, two-thirds to three-quarters of the way down the cannon.¹¹
• What they do. The medial splint carries the second and third carpal bones and the lateral one the fourth.¹ The 2006 surgical text Equine Surgery warns that removing more than half of the medial forelimb splint "can lead to carpometacarpal instability and lameness because MCII bears a substantial axial load."² The two splints and the cannon form the groove that houses the suspensory ligament, the elastic band that keeps the fetlock from collapsing under load.¹ A 2023 case report calls their fractures "the most common fractures in horses,"³ and a torn interosseous ligament is the "splints" of the young horse in training.⁴
• What they are not. They are not useless, and nobody who made the vestige argument at full strength said they were. Coyne 2009 calls them "true vestigial features" in the same parenthesis that names the disease they cause.⁵ A 2005 Zurich anatomy paper opens with "the vestigial metacarpal and metatarsal bones, so called splint bones".⁶
• The embryo. In 2020 Kavanagh, Bailey and Sears sectioned horse embryos of 29 to 35 days and found five digit condensations, with the splints forming by "fusions of the two anterior digits I and II and the two posterior digits IV and V."⁷ Coyne's 2009 statement that horse embryos begin with three toes is wrong.
• Extra-toed horses. O. C. Marsh examined the extra toes of living horses in 1892 and found that they grow from the splint bones, sit on the inner side of the fore foot, and were "supported in the usual manner" by their own carpal bone wherever the carpus could be inspected.⁸ A 2007 case report of a Berber colt found the same: three phalanges on the end of the second metacarpal and "well-developed first carpal bones."⁹
• The creationist replies. Sarfati (1999), Holt (2008), Menton (2014) and Bergman (2024) argue that the splints have functions.¹⁰¹¹ That answers the museum-sign version of the claim, and it answers it correctly. None of them says why a reduced second metacarpal should carry the second carpal bone, form from a fused pair of embryonic digits, or regrow into a hoofed toe with its own carpal bone. Holt and Bergman both call the extra toes a variation within the horse kind, which concedes reduction from a toed state and moves the whole dispute to the fossil series.¹¹¹²

Verdict: the splint bones are working parts of the horse's leg and reduced second and fourth digits at the same time.  Function is real and settles nothing.  Homology and reduction are settled by the anatomy alone, and the embryo and the extra-toed horses add two lines of evidence that a "designed that way" reading must explain and has not.  What the splints cannot do on their own is prove that the reduction happened in a lineage descended from three-toed and four-toed animals; that is the fossil series' job, and the sibling article prices it.

Part 1: What the splint bones are

A horse stands on one toe per leg.  The long bone between the knee, which anatomists call the carpus, and the fetlock is the third metacarpal, the cannon bone, and it carries the single digit.  Down the back of it, one on each side, run the two splint bones.  Sisson's Anatomy of the Domestic Animals of 1914 states it plainly: "Three metacarpal bones (Ossa metacarpalia) are present in the horse.  Of these, only one, the third or large metacarpal bone, is fully developed and carries a digit; the other two, the second and fourth, are much reduced, and are commonly called the small metacarpal or 'splint' bones."¹

Each splint has a head, a shaft and a free end.  The shaft "is three-sided and tapers to the distal end," and "is attached to the large metacarpal bone by an interosseous ligament, except near the distal end."¹  The head is the working end: "In the case of the medial bone it usually bears two facets above which support the second and third carpal bones, while the lateral bone has here a single facet for articulation with the fourth carpal bone."¹  The free end is "usually a small nodule, which projects to a variable extent in different subjects, and is easily felt in the living animal.  It is situated two-thirds to three-fourths of the way down the region."¹  Horsemen call it the button.  Sisson's small-print note on that button is the sentence the whole argument turns on: "The distal end is very variable in size and may be a mere point; in other cases, especially in large draft horses, it may present a prolongation which is regarded as the vestige of the digital skeleton."¹

The hind leg matches.  The small metatarsals "are a little longer than the corresponding metacarpals," the lateral one "is relatively massive, especially in its upper part" and meets the fourth tarsal bone, and the medial one "bears two facets above for the first and second tarsals, and sometimes one for the third tarsal."¹  In the joint chapter Sisson adds that the splints and cannon "are closely united by an interosseous metacarpal ligament (Lig. interosseum metacarpi), which often undergoes more or less extensive ossification."¹  The splints ossify from two centres, "Their distal ends are cartilaginous at birth," and "Fusion of the middle part of the shaft with the large metacarpal bone is common."¹

Nothing in the 1914 description has changed.  Dyce, Sack and Wensing's Textbook of Veterinary Anatomy (third edition, 2002) says the splints "are much reduced in size.  Each has a small proximal base that continues into a tapering shaft.  In young animals the splint and cannon bones are joined by fibrous tissue; this generally later ossifies and the upper parts of the shafts are then fused together."¹³

One piece of folklore the modern text corrects: the button is not a compressed toe.  Dyce writes that the splints "end in nodules, but the assumption that these incorporate greatly reduced elements of all three phalanges of the lost digits is unfounded."¹³  What is reduced is the metacarpal itself; the phalanges are gone.  Vincelette and colleagues put it the same way in 2023: horses "retain proximal remnants of the second and fourth metapodials, although the associated phalanges have been entirely lost.  These side metapodials are quite robust and form part of the support of the carpus and tarsus."¹⁴

Part 2: What they do, and what happens when they break

Load from the carpus.  The head of each splint is part of the floor the carpal bones rest on.  Dyce: "The carpus is mainly supported by the cannon bone but also makes contact with the bases of the splint bones.  Indeed, so large a part of the second carpal bone rests on the second metacarpal that it may tend to drive that bone away from its larger neighbor, inducing the painful acute inflammation mentioned later."¹³  Adams adds a structural point: "If one views a cross section of the metacarpus, it becomes obvious that the splint bones provide considerable support to the third metacarpal bone, since the dorsal cortex of the third metacarpal bone is much thicker than its palmar surface."⁴  The surgical text is the most explicit: "Prognosis after larger distal fragments of splint bones are removed depends on the size of the fragment removed as well as the specific bone involved.  The entire MTIV can be removed with good results, but removal of more than 50% of MCII can lead to carpometacarpal instability and lameness because MCII bears a substantial axial load.  Therefore, an effort should be made to repair rather than remove fractures of the proximal half of MCII."²  The same chapter says a page later that the load is small in absolute terms: "Fortunately, the limited weightbearing of the splint bone allows successful management when there is contamination and sepsis."²  Both are true at once.  The splints carry a minority of the load, and the medial front one carries enough that the carpus fails without it.

No one has measured the fraction.  Those statements are surgical experience and cross-sectional geometry, not strain gauges, and neither "substantial" nor "limited" has a number behind it.

The suspensory ligament.  The three metacarpals form a groove, and in it lies the suspensory ligament, which Sisson describes as "a wide, thick band" that "is attached above to the proximal part of the posterior surface of the large metacarpal bone and to the distal row of carpal bones,"¹ and whose "principal function is to support the fetlock, i. e., to prevent excessive dorsal flexion of the joint when the weight is put on the limb."¹  So the splints wall the ligament in; they are not its anchor, and the creationist phrase "attachment points for important muscles" needs care.  Sisson's muscle chapter is blunt: "The five muscles of the metacarpus and digit are either reduced to vestiges or modified greatly in structure."  The two small interossei that do arise from the splints, one from each, are "very small muscles, each of which arises from the corresponding small metacarpal bone near its proximal extremity, and is provided with a delicate tendon which is usually lost in the fascia at the fetlock (Fig. 240).  They have no appreciable action."¹  What the splints do anchor is ligament and fascia: in the hock, "Paired collateral ligaments extend from the malleoli to the corresponding splint bones and may be palpated along their whole lengths,"¹³ and a 2005 Zurich dissection of five cadaver limbs found the palmar and plantar fascia "a very stable structure, which connects the splint bone with its environment."⁶

"Splints."  The interosseous ligament that binds splint to cannon is the site of the commonest lameness of the young horse.  Adams: "A true splint refers to a sprain or tear of the interosseous ligament.  The resultant enlargement is most frequently observed 6 to 7 cm below the carpus on the medial side at the junction of the second and third metacarpal bones."⁴  It "usually affects young horses and involves the proximal one third of the medial splint bone."⁴  Dyce ties the site to the load: the condition "is more common at the medial intermetacarpal joint" because the second carpal bone bears on the second metacarpal and tends to drive it off the cannon.¹³  The bone that carries the most load is the one that gets the disease.

Fractures.  "Fractures of the splint bones are the most common fractures in horses," a 2023 case report opens.³  They are "most commonly located at the distal third."⁴  Verschooten, Gasthuys and De Moor in Ghent studied 87 horses with 114 splint fractures in 1984, found suspensory desmitis in "at least 70 per cent of horses," and reported that "more than 80 per cent of the fractures healed spontaneously.  Non-union fractures were not painful and did not cause lameness."¹⁵

What happens when the bone is taken away is the cleanest test of what it does.  The lower half is disposable.  Harrison, May and Edwards removed the distal splint in 9 of 26 open-fracture cases "in which the attachment of the splint bone to the cannon bone via the interosseous ligament did not provide adequate stability," and 25 of the 26 horses returned to soundness.¹⁶  The lateral hind splint is wholly disposable: "The entire MTIV can be removed with good results."²  In 2023 a 17-year-old Icelandic mare had her whole fourth metacarpal removed after a fracture, "reached her previous level of performance as a leisure horse" and "is still free of lameness 6 years after the incident."³  The upper half of the medial front splint is not disposable, which is why proximal fractures are plated rather than excised, and why they are "relatively rare and are challenging to treat."¹⁷

Part 3: The embryo, the relatives, and the five-digit dispute

The relatives.  The living perissodactyls are the horse, the tapir and the rhinoceros, and Marsh in 1892 set their feet side by side: "The tapir has the most primitive feet of existing perissodactyles, and most like those of the early predecessors of the horse," with "four toes in the front foot" and three behind; "The living rhinoceros has feet of the same general type, but less primitive," on three toes, "while the modern horse, when normal, has the extreme specialization of the monodactyle foot."⁸  In the tapir and rhinoceros the second and fourth digits are complete, with metapodial, three phalanges and a hoof; Solounias notes that "these taxa both have digits II and IV fully expressed."¹⁸  In the horse the same two bones are present, in the same position, with the same carpal articulations, and stop two-thirds of the way down.  That is what "reduced homologue" means, and no creationist source disputes the correspondence.

The embryo.  Coyne wrote in 2009 that "horse embryos begin development with three toes, which grow at equal rates.  Later, however, the middle toe begins to grow faster than the other two, which at birth are left as thin 'splint bones' along either side of the leg."⁵  That was the published account and it is wrong.  Kavanagh, Bailey and Sears obtained embryos from mares bred by timed artificial insemination at North Carolina State, serially sectioned the feet of five embryos aged about 29 to 35 days, and found that "early Equus embryos go through a stage with five digit condensations, and that the flanking splint metapodials result from fusions of the two anterior digits I and II and the two posterior digits IV and V, in a striking parallel between ontogeny and phylogeny."⁷  In the oldest embryos "the two anterior and two posterior digit condensations in these embryos are observed fusing (i.e. the merging of DI with DII, and DIV with DV), clearly showing that cell lineages from all five metacarpals persist in the older horse FL and that DI and DV are not absent or solely carved away as previously thought."⁷  The middle toe wins by growing, not by starting alone: "the central toe only emerges as dominant as a consequence of accelerated growth, development and later apoptosis relative to the other digits after the initiation of the five digit condensations."⁷  The hind foot runs ahead of the fore, as it does in the fossil series: "In one of our Equus embryos (Embryo 1), the HL lateral digits are fully fused while the FL digits remain distally separated."⁷

Five embryos is a small sample, the ages are estimates, and the fusions are read from stained sections rather than cell-lineage tracing; the authors say the earliest patterning stage "begs further detail from additional specimens."⁷  With those limits, the result stands: the horse builds five digits and merges four of them into two rods.  Menton's 2014 chapter for Answers in Genesis accepts the five-digit start ("Horses have five digits while developing as an embryo, but generally all but one (the third digit) is absorbed before birth"),¹⁹ so on this point the two sides do not disagree about the fact, only about what it means.

The five-digit dispute.  In 2018 Solounias and six colleagues proposed something stronger: that "all five digits are partially present in the modern adult forelimb."¹⁸  Their evidence is of three kinds: textured ridges on the palmar face of each splint, which they read as "surface expressions of metacarpals I and V, digits previously believed to be lost in these taxa";¹⁸ the two-lobed frog in the sole of the hoof, "newly theorized to represent a partial differentiation of the distal parts of digits II and IV";¹⁸ and the wings of the coffin bone and the hoof cartilages as digits I and V.  Bergman quotes it as evidence that the toes "were not lost, but are simply vestigial," which "supports the degeneration theory."¹²  It is not written that way.  The same paper treats the reduction as adaptive, citing Christine Janis for the view that "monodactyly evolves to allow the trot gait characteristic of the modern horse."¹⁸

The paper then split down the middle.  Kavanagh's embryos supported its proximal half: the ridges sit where the fused digit I and V material sits.  They rejected its distal half: "No evidence of any mesenchymal anomalies or distal remnant condensations was observed in the FL or HL that might account for digit cartilages contributing to the hoof elements (the frog), as hypothesized in Solounias et al."⁷  Vincelette's 2023 rebuttal in the same journal reached the same verdict: the proximal claim "is confirmed by the developmental evidence provided in Kavanagh et al.  However, their contention that distal remains of digits II and IV intersect to form the V-shaped frog on the posterior portion of the hoof, and that distal remains of digits I and V form the plantar processes plus collateral cartilages of the ungual phalanx, is shown here to lack support."¹⁴  Vincelette's own evidence is a frog impression in the tracks of three-toed Hipparion, which had full side toes and so could not have built its frog from them, and polydactyl horses whose extra toes carry frogs of their own.¹⁴

What survives is this.  All five digits are present at 30 days; I and II fuse into the medial splint and IV and V into the lateral; the distal parts of all four are removed; nothing of them is in the hoof.  That is the best-documented digit reduction in any living mammal, and the creationist literature has not engaged it.

Part 4: Horses with extra toes

Marsh's 1892 paper in the American Journal of Science dissected them.  His premise is the one this article started from: "It is not so well known, however, that the latter are the remnants of two other toes possessed by ancestors of the horse," and "One or more of these splint bones may become enlarged below, and support phalanges, forming another digit beside the main one, either beneath the skin, or, more commonly, developing into a small, external toe with hoof."⁸  He found the condition "much more frequent than is generally supposed,"⁸ and patterned: "This extra toe is almost invariably on the inner side of the main digit, and usually on the fore foot."⁸  The decisive observation is about the wrist: "In every specimen examined, where the carpal or tarsal series of bones were preserved, and open to inspection, the extra digits were supported in the usual manner," and "none of the extra toes examined represents the fifth digit."⁸  Marsh knew the rival explanation, duplication of the main toe "like the occasional sixth finger of the human hand," and that it "was long held by writers on the subject."⁸  His conclusion: "All the examples of polydactylism in the horse which the writer has had opportunity to examine critically are best explained by atavism, and many of them admit no other solution."⁸

Gould's 1983 essay on the same paper adds the later count: "A later, and particularly thorough, German monograph of 1918 concluded that about two-thirds of horses with extra toes had simply duplicated the functional third digit, while about one-third had resuscitated an ancestral feature by developing the vestigial splints of their second or fourth toe into complete, hoofed digits."²⁰

The modern case reports match Marsh.  Carstanjen, Abitbol and Desbois described a seven-month-old Berber colt at Alfort in 2007.  On the right fore, radiographs "revealed a supernumerary digit composed of proximal, middle and distal phalanges at the distal aspect of metacarpus (MC) II (Fig. 2) as well as the development of the first carpal bone"; on the left, "a supernumerary digit composed of one phalanx at the distal aspect of MC II (Fig. 3) and a well-developed first carpal bone (Fig. 4)."⁹  In the normal horse the first carpal bone is a pea: Sisson says it "appears to be absent on both sides in about half of the cases; in a good many subjects it is present on one side only."¹  The Alfort authors summarise the older literature: "About 80% of polydactyl cases show a supernumerary digit at the medial aspect of the forelimb.  In one study of 100 horses, 60% of the animals had one supernumerary digit, 24% bilateral involvement and 15% had this anomaly in all four limbs," and they give the standard three-way classification, in which teratogenic extra digits come from a split of the main toe, "whereas the splint bones are often normal," while "In the developmental or atavistic form, the digit usually derives from the distal end of the splint bone."⁹  Vincelette's summary is the fair one: polydactyly "is not necessarily a reversion to an ancestral state," but "in some cases, the supernumerary digit has its own carpal articulation and innervation and bears a resemblance to the ancestral condition," and in at least one case "the supernumerary digits also have a frog."¹⁴

The evolutionary reading is Marsh's: the splint's own developmental program, normally stopped at the button, occasionally runs on to phalanges and hoof, and the digit that reappears is the second, the last side toe to go in the fossil series, complete with the carpal bone that once supported it.  The creationist reading, from Holt in 2008, is that the extra toes show "a totally equine skeleton had extra long splint bones supporting its fossilized leg that at one time supported a member of the equine kind.  The modern equine's regressing splints suggest only that the genetic material for extended splints has been lost."¹¹  Bergman in 2024: "Modern horses have had 3 hooves at least until the 20th century and they are just as much a horse as the 1-hooved majority.  This adaptation is thus clearly a microevolutionary change, aka, a variation within their Genesis kind."¹²  Read carefully, both concede the reduced-homologue claim in full: the splints are "regressing" from toes that "at one time" were there.  What they deny is that the toed ancestor lies outside the horse kind.  That is a claim about the fossil series, and it is priced in the sibling article.

Part 5: The test, and the creationist literature in its own words

The parent's three steps go quickly here.  Homology: the second and fourth metacarpals of the horse sit where the tapir's do, carry the same carpal bones, and arise in the embryo from the same condensations, fused with the first and fifth.  Reduction: Sisson's "much reduced," the missing phalanges, and a muscle chapter in which the digit's own muscles are "reduced to vestiges."  Ancestral distribution: on the perissodactyl tree, three or four complete toes are the condition of every branch but one, and the fossil equids fill the gap in order.  McHorse, Biewener and Pierce showed in 2017 that the side toes of three-toed fossil horses were load-bearing, not passengers: "without accounting for side digits, three-toed horses as late as Parahippus would have experienced physiologically untenable bone stresses.  Conversely, when side digits are modelled as load-bearing, species at the base of the horse radiation through Equus probably maintained a similar safety factor to fracture stress."²¹  The head of the splint carries load because it is the retained part of a bone that carried load.  Function is not the alternative to vestigiality here; it is the continuity.

Now the creationist literature, in its own words.  Jonathan Sarfati in Creation in 1999: "Many evolutionists claim that the horse's splint bones in their legs (see diagram right) are vestigial, that is, useless leftovers from its alleged evolutionary past."¹⁰  His answer: "In particular, the horse's splint bones serve several important functions.  They strengthen the leg and foot bones, very important because of the enormous stress that galloping puts on the legs.  They also provide attachment points for important muscles.  And they form a protective groove that houses the suspensory ligament, a vital elastic brace that supports the horse's weight as it walks."¹⁰  The sentence is sourced to Bergman and Howe's 1990 book, page 77, which this project has not been able to obtain.²²  Rebekah Holt in 2008 quoted the farrier Doug Butler: "The function of the splint bones is to protect the tendons and ligaments and especially the blood vessels and nerves which pass down the back of the leg.  They also provide a greater bearing surface for weight by supporting a portion of the carpal bones of the knee joint.  They are necessary, not vestigial."¹¹  David Menton, for Answers in Genesis in 2014: "Vestiges of the second and third metacarpal (and metatarsal) bones are visible in the modern horse as the splint bones."¹⁹  Jerry Bergman in 2024 read Solounias as showing the toes "were not lost, but are simply vestigial," which "supports the degeneration theory."¹²

Where these are right: the useless claim is unprovable and, for the splints, false, and the anatomy above says so at length.  Sarfati's strengthening and groove claims are correct; Butler's carpal-support claim is correct and is the strongest of the set.  Where they go wrong is the target.  Sisson in 1914, Coyne in 2009 and the Zurich anatomists in 2005 call the splints vestiges in the same paragraphs that describe their facets, their groove and their fascia; "vestigial" in every one of those sources means reduced, not useless.  Two details need correcting on the creationist side.  "Attachment points for important muscles" is wrong on Sisson's account: the only muscles arising from the splints are two interossei with "no appreciable action."¹  And Menton's "second and third" should read second and fourth; the third metacarpal is the cannon.  Bergman's use of Solounias is answered in Part 3.

What a creationist would need to say instead is one of two things.  Either the horse was made one-toed, with two reduced side metacarpals as a designed carpal support, in which case the five embryonic condensations, the fusion of I with II and IV with V, the button "regarded as the vestige of the digital skeleton," and the extra toes that grow from it with their own carpal bone are all coincidences of design that the literature has not addressed.  Or the horse kind once had side toes and lost them, which is what Holt and Bergman both write, in which case the reduced-homologue claim is conceded and the only live question is where the kind's boundary falls in the fossil series.  No creationist source found for this article takes the first road explicitly, and the second is the sibling article's subject.

Objections and responses

"They carry load and house the suspensory ligament, so they are not vestigial."  They carry load, and the medial front one carries enough that a surgeon will plate rather than remove it.²  This answers the claim that they are useless, which Sisson, Marsh, Coyne and the veterinary anatomists never made.  The claim they did make is that the splints are the second and fourth metacarpals, reduced.  A load-bearing head is what a reduced load-bearing bone would keep.

"The extra toes are duplications, like a sixth finger, not reversions."  Most are.  Gould's count is two-thirds duplication.²⁰  The argument rests on the other third: toes that grow from the end of the splint, on the medial side, with their own carpal bone and their own frog, in the position of the second digit that three-toed fossil horses carried last.⁸⁹¹⁴  A duplicated third toe does not bring a trapezium with it.

"Five embryonic condensations are just the shared tetrapod program.  A designer reusing one program would give the same result."  Perhaps, but that reading grants the homology, which is step one of the test, and it makes the rest a coincidence: that the five buds fuse into exactly the two rods the fossils show, that the hind foot fuses first as the fossil hind foot lost its toes first,⁷ and that the adult button is what Sisson called the vestige of the digital skeleton.  Common descent predicts each of those; a reused program predicts none of them in particular.

"Solounias showed all five digits are still there, so nothing was lost; that is degeneration, not evolution."  Solounias's proximal claim survived and his distal claim did not,⁷¹⁴ and neither half says what Bergman says it says.  Fused rudiments of five metacarpals with no phalanges on four of them is reduction by any definition, and the paper that reported it calls the change adaptive.¹⁸  "Degeneration" and "reduced homologue" are the same observation under two names.

"This article concedes the case."  It concedes homology and reduction, which Holt and Bergman also concede when they call the extra toes a variation within the horse kind.  The disagreement that remains is whether a three-toed or four-toed animal in the fossil series is the same kind as Equus, and the splint bones cannot answer it.  Read The Horse Series for that.

Issues and unanswered questions

• No published in-vivo measurement of the load through a splint bone was found. "Substantial axial load" and "limited weightbearing" are surgical judgement, and a strain-gauge study on the proximal MCII would settle how much the carpus depends on it.
• The frequency of polydactyly in living horses is unquantified. Marsh's "much more frequent than is generally supposed" carries no number, the 100-horse series is Lindemann's 1909 Leipzig dissertation, read here only through Carstanjen, and Gould's 1918 German monograph was not identified.
• The atavistic and teratogenic forms of polydactyly are separated by radiographs and judgement, and Carstanjen's authors call the classification "controversial."⁹ The Alfort owner's report that the extra foot "was not present at birth" is unexplained.⁹
• Kavanagh's five embryos are a small sample, with estimated ages and no lineage tracing; nobody has repeated the study.
• Stanek and Hantak's 1986 case of a colt and its dam, the one report of polydactyly running in a family, is known here only through Vincelette and Carstanjen; it has no abstract in PubMed and was not obtained.
• Bergman and Howe 1990, page 77, the source Sarfati cites for the function list, was not obtained; it may contain arguments not represented here.
• No creationist source found for this article addresses Kavanagh 2020 or the carpal bones of polydactyl horses.

Verdict

The splint bones are the second and fourth metacarpals and metatarsals of the horse, reduced to tapering rods that keep their carpal facets, wall in the suspensory ligament, tear their ligament in two-year-olds and are among the commonest fractures in the animal.  None of that is in dispute, and none of it was disputed by the anatomists who called them vestiges.  On homology and reduction the case is closed: the tapir has the same two bones with toes on them, the embryo builds five digits and fuses four of them into these two, and about a third of extra-toed horses regrow a second digit from the splint, with the carpal bone to carry it.  The creationist literature has answered a claim of uselessness that the primary sources never made, and its own "variation within the kind" reading concedes reduction from a toed state.  Whether that toed state lies inside or outside the horse kind is a question about the fossil series, not about the splints, and that is where the argument now lives.

Sources