Sunday, June 15, 2014

What Does T. rex Say?

T. rex holotype specimen. Photo by Scott Robert Anseimo, CC BY-SA 3.0.
It's an iconic scene in every dinosaur movie: the huge, conquering carnivorous theropod rears back and lets out a terrifying bellow. Sound effects artists spend huge amounts of time sampling vocalizations from various animals to create just the right mix to create an unfamiliar, otherworldly roar. And, of course, everybody knows that pterodactyls let out harsh, echoing, prehistoric sounding screeches.

But how close to reality are these sounds? Do we have any ways of using science to figure out what dinosaurs and other stem-birds may have sounded like? Do we have evidence that they made sounds at all?

Take a look at the Vulture article linked above. It lists some of the amazing places sound effects artists went to create the sounds for the dinosaur movie Jurassic Park in 1993. The iconic T. rex roar was created by playing with the speed and frequency of elephant and dog sounds. The raptor sounds were created using tortoises, horses, and geese. Only one of these sources is an animal anywhere close to being related to dinosaurs--the goose. Many of the other sounds are re-mixed from various mammals, the kinds of sounds  we expect large predators (which today are almost all mammalian) to have, thus increasing the scare factor for audiences.

In reality, mammalian or even tortoise sounds are probably not the best analogue for real-life dinosaur sounds. But then what are? The only real evidence for stem-bird vocalizations in fossils comes from two sources: good old phylogenetic bracketing, and correlations between bone structure and soft tissue vocal chords. Phil Senter reviewed this evidence in his 2009 paper on evidence for vocalization in prehistoric animals.

Many groups of animals today can make noise using vocal chords of some kind, and most of these do so using some kind of larynx or larynx-like vocal organ. Most groups of mammals vocalize, so the prince of a larynx is probably ancestral for our own major group, though how far back down the stem-mammal lineage the larynx might go is hard to say. One line of evidence cited by Senter is the presence of a tympanic ears, which would be necessary to hear any complex vocal communication being made. The earliest stem-mammals lacked ears, which don't appear until almost the mammalian crown group, though the grade leading up to the crown may have been able to hear some low-frequency airborne sounds.

The almost random distribution of tympanic ears and the larynx in modern tetrapods indicates that these features evolved multiple times, for example in frogs, some salamanders, mammals, turtles, gekkotan lizards, and crocodiles. This is supported by the fact that birds lack the larynx entirely, and vocalize with an entirely different organ called the syrinx. While it is possible that the ancestors of birds had a larynx and it was later replaced with the syrinx, there is currently no evidence to support or test this possibility, and it should be assumed that stem-birds never had a larynx.

Can we tell how far down the bird family tree the syrinx evolved? It turns out that the presence of a sound-making syrinx is strongly tied with the presence of a clavicular air sac, without which the syrinx could not function (Olsen & Joseph, 2011). Like other air sacs, the clavicular sac creates hollow spaces in the bones of the forelimb and shoulder girdle. There are three known stem-bird clades that have such air sacs: Pterosauria, Aerosteon, and Ornithothoraces. The latter, of course, is the one that includes birds, so the air sacs in non-avian ornithothoracines are probably homologous with those of modern birds. Aerosteon and pterosaurs, though, are problematic. Aerosteon are bracketed on both sides by theropods known to definitely lack clavicular air sacs, so the presence in the former taxon is probably an independent evolution possibly unrelated to vocalization. The same goes for pterosaurs, which may have independently evolved air sacs to help cope with flight, and independently developed the air sacs in their forelimbs, possibly without any relation to a syrinx (Senter, 2009).

Does this mean T. rex and her dinosaurian buddies were all silent? The answer is no, because animals can make plenty of sounds by means other than a larynx or syrinx, just not the mammal- or bird-like sounds we are used to associating with them in pop culture. Hisses, drumming sounds, clicking or rattling sounds can all be produced just by moving air and not vocalizing. Non-ornithothoracine stem-birds could also have made physical sounds, like whip-cracking tails in diplodocids, beak-clacking, jaw-grinding, water-slapping, etc. Some birds can also make low booming and drumming calls using only passages through the trachea and cervical air sac, like emu and cassowary (Olsen & Joseph, 2011).

What about the famous hooting horns and crests of lambeosaurines? Senter noted that even in modern animals, the presence of resonating chambers don't always correlate with the presence of vocal chords. Some snakes, for example, have massive resonating chambers in their skulls which amplify and modify the sound of their hiss. It's entirely plausible, Senter argued, and more in line with the other evidence, to think that Parasaurolophus used their crests to project loud hissing or drumming sounds, rather than trumpeting (Senter, 2009). However, it's also possible that the complex internal, trumpet-like tubes of such crests evolved because these animals lacked vocal chords. After all, a trumpet player doesn't vocalize into his instrument, he simply blows air through, and the internal "anatomy" of the trumpet create the musical sound. I have to disagree with Senter here and suggest that lambeosaurines did hoot and trumpet, but were probably one-note instruments, unless some kind of weird soft tissue allowed them to hit different "keys" and modify the sound. Either way, no syrinx or larynx should be required for this to work.

It needs to be noted that the larynx evolved many times in tetrapods, and it's entirely possible that some stem-bird lineages also independently evolved such an organ. But absent any kind of testable, direct evidence, or unusual sound-related structures like the ones found in lambeosaurines, any good scientist needs to presume that most of them could not vocalize the way depicted in movies.

So how did these animals communicate if their were largely silent and limited in vocal repertoire? Stem-birds, more than possibly any other group of tetrapods, are well-known for outlandishly flamboyant visual display structures. Huge, flashy crests, plates, spikes, sails, and frills abound in nearly every stem-bird group. The idea that these animals were very limited in audible communication may help explain why they developed such a dazzling array of visual communication devices. It is only in the ornithothoracines that we see direct evidence for acoustic communication and, maybe not coincidentally, this is the first group of stem-birds to become fully arboreal. For very small species living mostly hidden from each other in the dense leaves of trees, visual communication would become more difficult, and there would be selective pressure for alternate means of communication, such as sound, to evolve.

Bird calls may have developed initially to keep in touch between animals that had become hidden from each there in the trees and could no longer regularly maintain visual signals. Like songbirds today, which use visual but mainly vocal communication to find and attract mates, basal ornithothoracine protobirds may have become the first stem-birds to call to each other through the foliage.

So, what did T. rex say? It may have hissed and boomed like crocodiles, snakes, and emus. It may have been the strong, mostly-silent type. But it probably couldn't roar. Except, of course, in the next Jurassic Park film. Movie producers will never let science get in the way of a good scare!


  1. Interesting post. I reached similar conclusions in two old posts:

  2. Ah good food for thought! I'm actually working on some prehistoric monster sound design right now (albeit for entirely fictitious creatures!)...

    (see here: )

    As an artist I'm inclined to argue that it's fair to give a pretty high degree of creative license to sound designers in blending a variety of sounds in order to come up with something different enough that it sounds like it could've been plausibly emitted from a vast vast group of extinct animals whose soft tissue is largely unknown… but I am just an artist, and i'm open to the possibility that I'm biased and stubborn...

    Something worth mentioning from a sound design perspective is that, strictly sonically speaking (as in how these various sounds sound), the tonal range and texture of many gutteral aggressive roars coming from crocs, ratites, and big mammals is often very similar… Close your eyes and listen to an elephant roar, or a big crocodile, or an attacking cassowary and many of the sounds are sort of similar even if they are made by entirely different structures.

    And (again) maybe I'm just a stubborn creative speaking from my gut, but I don't think a phylogenetic analysis is enough to conclude that "it should be assumed that stem birds never had a larynx… (or presumably ANY other vocalizing structures?)" I just don't think it's a fair assessment of the soft tissue possibilities when considering groups with such stunning diversity as was present in extinct mesozoic groups. In addition to the repeated evolution of the larynx and tympanum, consider, if you will, the repeated and disparate evolution of oviparity, viviparity and oviviparity in extant squamates, and what little evidence would be left in the fossil record of their apparent reproductive pliability… Also, what's the deal with gecko vocals? Is there any bony evidence of their talkin boxes? (not rhetorical, i have not idea. I do know however, they make a wide variety of strange noises some of which sound like birds, dogs, rodents, or monsters).

    Also when you consider that several of the groups you compared mesozoic theropods to are flight adapted (which requires a crazy level of skeletal and soft tissue adaptation) is the fact that large mesozoic theropods definitely lacked clavicular air sacs evidence AGAINST them having another means of complex vocalization? Birds certainly don't have the big muscular necks, barrel chest and and gaping skull fenestra that say, t-rexes did…

    With regards to your assessment of communication via visual display vs. aural, without good genetic and physiological evidence I don't think it's remotely reasonable to assume that because mesozoic dinosaurs have a lot of wacky skeletal features that 1) they're all definitely visual displays and 2) it would be easier to evolve and grow those metabolically expensive bone structures than it would be to evolve (and lose and re-evolve) various soft tissue sound makers. Extant birds are certainly doing a lot of BOTH visual and aural displaying often in dense forests, deserts, oceans, cliff faces and ALL of that diversity evolved in the last 65 million, much of it sooner. Consider the various balloon necked grouses, casque headed hornbills, and mannequin birds all of whom COMBINE visual and aural displays, features of which simply wouldn't be detectable in the fossil record… The competition for territory and mates is FIERCE!!!

    If this was the only animal we could observe would phylogenetic analysis support the claim that all birds with structurally similar wings made communicative sound with their feathers?

  3. What does a T. rex say? Possibly an earsplitting roar if modern birds are any indication. I've read the cited paper, and I have several problems with his arguments, and I think there are some factual errors that need to be addressed.

    The larynx is not randomly distributed across Tetrapoda: It’s a symplesiomorphy among amniotes (it can be found in lungfish), and more importantly for this discussion, birds DO have a larynx. I think people tend to confuse the larynx with vocal cords; they are not the same thing. Most tetrapods possess a larynx, parts of which have been independently modified into vocal cords or flaps in various lineages such as frogs, turtles, geckos, crocodilians, and mammals.

    Since modern birds possess a larynx, as do crocodilians, we can be reasonably sure that most if not all archosaurs also possessed a larynx also.

    I also think you are confusing postcranial pneumatization with the presence of air sacs; these are not equivalent. During embryonic air sac ontogeny, the air sacs bud from the mesobronchi before invading cervical skeleton, followed by a pattern of pneumatization that appears to recapitulate the parallel evolution of the same among different ornithodiran lineages. Certainly, the clavicular air sac must have evolved before invading the surrounding bones! The presence of clavicular, cervical, and abdominal air sacs in pterosaurs and some theropods suggests that the basic avian bauplan might be a synapomorphy within Ornithodira, with only a few derived lineages evolving PSP in parallel. The parallel evolution of such a complex system is more parsimonious than the convergent evolution of the same.

    We cannot currently determine when the syrinx evolved, but if the clavicular air sac is a synapomorphy among ornithodirans, it's possible that the syrinx is also. Birds vocalize by using air pressure with the clavicular/interclavicular air sac to drive the lateral tympanic membranes medially and rostrally into the lumen of the bronchi. Exhaled air vibrates the LTM, producing sound, and in some birds, two sounds at the same time. If sauropods possessed a syrinx, their extremely long trachea would have resonated the sound, much like the trachea of whooping cranes and trumpeter swans, so the singing of the brachiosaurs in Jurassic Park might be more accurate than we realize. Since T. rex is bracketed between theropods that possessed clavicular air sacs, it’s quite possible that tyrannosaurs produced loud, deep roars for mating calls and territorial threats. Even if the syrinx is limited to just Theropoda or Saurischia, it’s still possible that ornithischians evolved laryngeal folds or flaps that allowed for phonation. Moreover, the convoluted tube crests of Parasaurolophus resemble the folded trachea of whooping cranes and those of some other birds.

    1. I agree with you on that one, I think that dinosaurs had larynxes, but they probably did not sound like they do in the movies.

  4. "After all, a trumpet player doesn't vocalize into his instrument, he simply blows air through, and the internal "anatomy" of the trumpet create the musical sound."

    No they do not. The trumpeter presses his or her lips together to produce a buzzing sound, much like how the LTM of birds produces sound when air pressure from the clavicular air sac presses the membranes together. If one is going to use a trumpet as an analog for lambeosaurine crests, then it would be an argument *for* vocal cords, not against!

    The larynx, which is an ancestral character of tetrapods, seems to show a propensity for being modified into a vocal organ. Considering the well-developed ears of archosaurs, it’s possible that the earliest archosaurs possessed simple, laryngeal folds, retained in modern crocodilians, which were later replaced by the more versatile syrinx-airsac system within the pan-avian lineage.

    All things considered, we simply do not know to what extent non-avian dinosaurs could vocalize, but I think we have enough information to at least say that vocalizing non-avian dinosaurs, or even pterosaurs, are at least *plausible*. I don’t think visual and sound effect artists should feel guilty when they give a T. rex a terrifying roar rather than a snake-like hiss.

    -Simeon Koning

  5. In addition to the complications others have brought up, it appears that birds do have a larynx and even use it to aid vocalization in at least some species.

    On a different note, it may be of interest that there are extant theropods that lack a syrinx (New World vultures) and tend to make demonic hissing sounds.

  6. Tyrannosaurus walked down the street.
    He opened his mouth and roared... Tweet! Tweet!
    -- BONE POEMS by Jeff Moss

  7. There is also the possibility that dinosaurs did not sound like any modern animals. I think that the dinosaurs had vocalizations that were unlike anything we can experience today.