Reader Comments

Post a new comment on this article

Animal species as possibly confounding factor?

Posted by Stephan_Matthiesen on 15 Feb 2014 at 15:05 GMT

I wonder if the species of the animal that is being depicted is a confounding factor and the analysis should be repeated for each species separately.

From casual observations it seems to me that elephants, for example, move their legs generally more slowly that horses, so that their gait pattern might be much easier to observe correctly. I would also imagine that people in different historic periods are interested in different animals. If the prehistoric images in your database tend to show more larger, slowly moving animals like elephants and giraffes than the modern images, then your finding might, to some degree, be explained not by better observing skills, but by the fact that different animals were being observed.

I may have overlooked the information, but it is not clear to me how the proportion of different animals varies with historic period in your database. It might also be helpful to perform the analysis for horses only rather than all quadrupeds (presumably horses are the most frequently shown animals).

No competing interests declared.

RE: Animal species as possibly confounding factor?

ghorvath replied to Stephan_Matthiesen on 26 Mar 2014 at 10:06 GMT

Dear Stephan Matthiesen,

thank you for your interesting comment.

(1) YOU WROTE: From casual observations it seems to me that elephants, for example, move their legs generally more slowly than horses, so that their gait pattern might be much easier to observe correctly. I would also imagine that people in different historic periods are interested in different animals. If the prehistoric images in your database tend to show more larger, slowly moving animals like elephants and giraffes than the modern images, then your finding might, to some degree, be explained not by better observing skills, but by the fact that different animals were being observed.

(1) ANSWER: According to the point A (Detailed analysis of prehistoric quadruped walking illustrations) of the Supporting Information and the Supplementary Figures S1-S35 of our paper, the statistics of quick and slow animals in prehistoric walking depictions is the following (see Appendix 1 below):

quick quadrupeds (28 items = 71.8%): 11 incorrect (39.3%) + 17 correct (60.7%)

slow quadrupeds  (11 items = 28.2%):  7 incorrect (63.6%) +  4 correct (36.4%)

Hence, only 28% of the analysed prehistoric depictions illustrated slow animals, and 72% depicted quick quadrupeds. Interestingly, only 39.3% of quick depictions was incorrect and 60.7% was correct, while 63.6% of slow depictins was incorrect and only 36.4% was correct.

From this we conclude the following: Our findings that the prehistoric quadruped depictions had the lowest error rate (46.2%) cannot be explained by the hypothesis that cavemen might have depicted predominantly slow animals, the leg attitudes of which could have been observed easier than those of quicker quadrupeds. Therefore, considering the prehistoric quadruped walking depictions, the animal species (slow or quick) was not a confounding factor.

---

(2) YOU WROTE: I may have overlooked the information, but it is not clear to me how the proportion of different animals
varies with historic period in your database. It might also be helpful to perform the analysis for horses only rather than
all quadrupeds (presumably horses are the most frequently shown animals). I wonder if the species of the animal that is being depicted is a confounding factor and the analysis should be repeated for each species separately.

(2) ANSWER: According to Appendix 2 (see below), modern quick (94.9%) and slow (5.1%) quadrupeds have an error rate of 67.9%, and 32.7%, respectively. Among quick quadrupeds, horses (88.6%) and other quick animals (6.3%) have an error rate of 69.7%, and 42.6%, respectively. Circa 95% of the studied modern quadruped walking depictions illustrated quick animals and only ca. 5% depicted slow quadrupeds. Circa 89% of the investigated modern walking quadruped depictions illustrated horses, and only ca. 6% depicted other quick animals.

Consequently, there was a clear tendency that the modern walking depictions of slow quadrupeds had a much lower error rate (ca. 33%) than that of quick quadrupeds (ca. 68%). Furthermore, the modern illustrations of horses had a higher error rate (ca. 70%) than that of other quick animals (ca. 43%).

With best wishes: Gabor Horvath (corresponding author)

********************************************
Department of Biological Physics
Physical Institute
Faculty of Natural Sciences
Eotvos University
H-1117 Budapest, Pazmany setany 1
Hungary
gh@arago.elte.hu
http://gh@arago.elte.hu
********************************************

APPENDIX 1: Prehistoric quadruped walking depictions (39 items)

QUICK QUADRUPEDS (horse, cow, bull, antelope, deer)

HORSES: 8 items (3 incorrect + 5 correct)
Figures S13 (cell Bf: incorrect), S14 (cell Hf: incorrect), S18 (cell Be: incorrect)
Figures S2 (cell Ca: correct), S2 (cell Cb: correct), S5 (cell Bb: correct), S16 (cell Bb: correct), S17 (cell Ba: correct)

COWS: 4 items (4 correct)
Figures S1: cow (cell Gf: correct), S8 (cell Bb: correct), S8 (cell Ba: correct), S26 (cell Hh: correct)

BULLS: 8 items (4 incorrect + 4 correct)
Figures S3 (cell Da: incorrect), S12 (cell Fh: incorrect), S15 (cell Dh: incorrect), S33 (cell Df: incorrect)
Figures S6 (cell Bb: correct), S11 (cell Ff: correct), S15 (cell Hh: correct), S27 (cell Hh: correct)

ANTELOPES: 5 items (4 incorrect + 1 correct)
Figures S9 (cell Bf: incorrect), S10 (cell Be: incorrect), S28 (cell Fb: incorrect), S29 (cell Fb: incorrect)
Figure S30 (cell Ff: correct)

DEERS: 3 items (3 correct)
Figures S20 (cell Fe: correct), S25 (cell Ba: correct), S31 (cell Hh: correct)

---

UNKNOWN SPECIES: 1 item (1 correct)
Figure S32 (cell Bb: correct)

---

SLOW QUADRUPEDS (giraffe, elephant, buffalo, rhinoceros, mammoth)

GIRAFFES: 3 items (1 incorrect + 2 correct)
Figures S4 (cell Fb: incorrect), S19 (cell Fe: correct), S22 (cell Bb: correct)

ELEPHANTS: 5 items (4 incorrect + 1 correct)
Figures S7 (cell Be: incorrect), S7 (cell Cf: incorrect), S7 (cell Bf: incorrect), S21 (cell Gd: incorrect)
Figure S35 (cell Cb: correct)

BUFFALO: 1 item (1 incorrect)
Figure S23 (cell Hb: incorrect)

RHINOCEROS: 1 item (1 correct)
Figure S24 (cell Cb: correct)

MAMMOTH: 1 item (1 incorrect)
Figure S34 (cell Bf: incorrect)

------------------------------------------------------------------------------------------------------------

APPENDIX 2: Modern quadruped walking depictions (961 items)

QUICK QUADRUPEDS: 912 items (94.9 % of the total 961) = 619 incorrect + 293 correct > error rate = 67.9 %

quick quadrupeds other than horse: 61 items (6.3 % of the total 961) = 26 incorrect + 35 correct > error rate = 42.6 %

horse: 851 items (88.6 % of the total 961) = 593 incorrect + 258 correct > error rate = 69.7 %
cow:  14 items (5 incorrect + 9 correct)
deer:  11 items (6 incorrect + 5 correct)
big cat:  8 items (2 incorrect + 6 correct)
dog: 5 items (2 incorrect + 3 correct)
moose:  5 items (4 incorrect + 1 correct)
goat:  3 items (1 incorrect + 2 correct)
zebra: 2 items (1 incorrect + 1 correct)
antelope:  2 items (0 incorrect + 2 correct)
sheep:  2 items (1 incorrect + 1 correct)
cat: 2 items (1 incorrect + 1 correct)
weasel: 2 items (1 incorrect + 1 correct)
coyote: 1 item (1 incorrect + 0 correct)
wolf: 1 item (0 incorrect + 1 correct)
donkey:  1 item (1 incorrect + 0 correct)
fox:  1 item (0 incorrect + 1 correct)
hyena: 1 item (0 incorrect + 1 correct)
 
---

SLOW QUADRUPEDS: 49 items (5.1 % of the total 961) = 16 incorrect + 33 correct > error rate = 32.7 %

camel: 17 items (6 incorrect + 11 correct)
bear: 17 items (5 incorrect + 12 correct)
buffalo:  8 items (3 incorrect + 5 correct)
elephant:  5 items (1 incorrect + 4 correct)
rhinoceros: 2 items (1 incorrect + 1 correct)

No competing interests declared.