Mandukasana

The Gheranda Samhita describes mandukasana as turning the toes back and bringing the knees forward, so that the practitioner sits low with the legs folded out to either side. Manduka is the frog, and the splayed legs with the feet turned behind are the animal seen from above.

The posture is one of the thirty two the text presents as useful in this world, out of the eighty four it calls best and the countless number it says exist. It is not in the Hatha Yoga Pradipika, whose fifteen postures are almost all cross legged seats. The frog belongs instead to the low kneeling and squatting group that the later compiler adds, alongside forms such as vajrasana and utkatasana, and that group is one of the clearest indications that the Gheranda Samhita drew on practice its predecessor did not record.

The Gheranda Samhita immediately follows it with uttana mandukasana, the stretched frog, which takes the same arrangement of the legs and raises the upper body with the arms behind the head. Paired entries of that kind appear only three or four times in the whole corpus, and each time the second posture is described as beginning from the first. The pairing is a small but genuine piece of evidence that the compiler was working from performed practice rather than assembling a list from written sources alone.

Counted across the classical corpus, animal names make up the largest single group in the posture vocabulary: serpent, eagle, cow, cockerel, tortoise, crocodile, frog, fish, peacock, lion, locust, camel, bull and tree stand beside a much smaller set named for people, for qualities and for objects. The Gheranda Samhita carries most of them, and the Hatha Yoga Pradipika comparatively few.

That distribution follows from the shapes each text records. The earlier manual is a book of seats, and a seat is hard to name after an animal because animals do not sit cross legged; its names accordingly tend to be qualities, the auspicious, the liberated, the accomplished. Once a compiler admits prone, kneeling and standing forms, resemblance becomes available as a naming principle, and the animals arrive with the shapes.

None of this implies that an animal posture is a way of taking on an animal's nature. The manuals do not say so, and the atlas does not import that reading from later popular writing. The names work by likeness, and they work well: a name that carries a picture is far easier to transmit orally than a description of where each limb goes, which matters in a tradition whose technical literature is terse verse intended to be memorised. Mallinson and Singleton's Roots of Yoga presents these lists in translation and makes the pattern easy to see across the corpus as a whole.