‘Wheels’ in nature
Escher was simultaneously wrong and amazingly right in a prophetical way. He was wrong, because several instances of wheel-like organisms or parts of organisms have subsequently been detected and discussed. He was right, because two of these natural examples correspond to an amazing degree to his fantasy-born animal. As the ‘Curl-up’ these real animals are segmented worm-shaped creatures that normally walk but, under certain conditions, form a wheel and start rolling. One of these is the stomatopod crustacean Nannosquilla decemspinosa Rathbun, 1910, which lives at the Pacific coast of Panama. If a wave washes it onto the sand beach, it rolls back to the water by means of backward somersaults and consecutive rolling by forming a wheel with its entire body (Caldwell, 1979; Full et al., 1993) (Fig. 3). The other example is the mother-of-pearl caterpillar Pleuroptya ruralis Scopoli, 1763 that also rolls away backwards by adopting a wheel shape with its body when it is attacked (Brackenbury, 1997). The major difference between the two real and Escher’s artificial animals is that the latter rolls forward whereas the caterpillar and the stomatopod roll backward with the tip of the tail ahead (Fig. 3). The prophetic view of the artist is nonetheless amazing.
One can add some more examples, such as the salamander Hydromantes platycephalus (Camp, 1916), or pangolins which curl themselves up and which have been observed to either passively or actively role away from predators (Tenaza, 1975; García-París and Deban, 1995). Enrolment is a widespread phenomenon and seen, for instance, in animals such as hedgehogs, armadillos, lizards (e.g., Cordylus cataphractus Boie 1828), amphibians (e.g., Taricha granulosa (Skilton, 1849), Echinotriton chinhaiensis (Chang, 1932) (see Brodie et al., 1984; Johnson and Brodie, 1975), isopods, myriapods, and even fossil trilobites (Fig. 4). The latter are the oldest examples so far, since enrolled trilobites have already been reported from the Cambrian (Bergström, 1973). This curl-up behaviour is primarily used for protection against a predator, but might also include an intended or unintended passive rolling if there is a slope.
Other animals and plants use their round or cylindrical body for passive rolling. The tumble weed, Corispermum hyssopifolium Linnaeus, 1753, which is well known to most of us from classical Western movies, falls into this category. This herb grows into a rounded form that becomes disconnected from the grounded roots and spreads its seeds via a rolling motion caused by the wind. Another prominent example is the spider Carparachne aureolava Lawrence, 1966 from the Namibian desert that can form a wheel by pulling in its legs and passively rolling down slopes (Henschel, 1990).The anomalan mole crab Hippa pacifica (Dana, 1852), which lives in the intertidal of sandy beaches in the tropics, shows a complex behaviour that helps this animal cope with the forces and turbulences of the breaking waves (Lastra et al., 2002). This behaviour includes lateral rolling of this almost cylindrically shaped animal moving down the slope of the beaches with the outgoing waves (own observation and pers. comm., M. Lastra, 2007). In all these cases, the whole animal or plant part undergoes the rolling movement and, accordingly, the similarity to a wheel on a cart is only given in part.
These examples of passively and actively rolling entire animals and plants lead to the problem of why wheels were not invented by nature as organs, e.g., an insect with six wheels instead of six legs. In fact, at the ultrastructural level a wheel-like organ is realised. It is found in the rotation of bacterial lagella showing a biochemical rotating motor at the lagella bases (Berg, 2003). However, bacteria do not roll on wheels in the strict sense. Furthermore, in the macroscopic biological world an organ with the shape or function of a wheel is not realised, not even similar to bacterial lagella.
There are two main reasons suggested for this. One is an external constraint specifying that under most natural conditions wheels are not advantageous (Gould, 1981; Walker, 1991; Meyer and Halbeisen, 2006) and indeed, everyone who ever left the smooth surface of streets and got stuck in mud or sand with a wheeled car or bike has experienced that this is a serious problem. However, this argument is only partially correct, because there are of course habitats where the ground is relatively smooth and hard and where the low friction resistance of a round surface might be advantageous (as is shown in the examples above).
Fig. 4. More than 500 million years of enrolled arthropods. Left: An Ordovician trilobite (Asaphus raniceps Dalman, 1827). Right: A terrestrial isopod (Armadillo officinalis Dumeril, 1816).
The second line of reasoning is based on internal constraints (Gould, 1981; Meyer and Halbeisen, 2006). According to this view, wheels need a certain functional autonomy because they have to freely rotate around an axis. Organs, however, grow and must be innervated and supplied with nutrients, oxygen etc. Hence, an organ cannot function as a wheel (Gould, 1981; Meyer and Halbeisen, 2006). Again, this view seems too restricted. One could imagine a secretion product, or dead cellular material such as hair or horn, that is produced and centred around an organic structure formed during development as a ring, and then some glands that secrete a lubricant. That this example is not too far fetched is exempliied by the mucous product of the clitellum of annelids such as earth worms, which is formed as a ring around the body and which loses contact with the region that produces it (Peters and Walldorf, 1986). Of course, wheels of that kind are just good for passive rolling, but one can think of indirect propulsion with other body structures such as tails. Nevertheless, it is a matter of fact that a true wheel as an organ is not present in nature.
However, do other kinds of wheels occur in nature? What we consider a wheel depends on the wheel’s definition, consciously or unconsciously used. If we define a wheel in the strict sense of Meyer and Halbeisen (2006), who state that a wheel is defined as a ‘round object rotating around a fixed axis’, then wheels are not realised in nature indeed. However, if we allow a more general definition by considering that the important element of wheel function is the use of the low friction resistance of round structures to carry a weight along a distance, then we do find wheels in nature. Perhaps we have to think along different lines, not so much about wheel organs but about wheel instruments or tools, or in other words, wheels produced and rolled by animals.
One such example is the perfectly round sand pellets produced by the mouthparts of the Australian sand bubbler crab Scopimera inflata A. Milne Edwards, 1873 (Fielder, 1970). These balls are the remains of feeding small organic particles in the sand of subtropical and tropical beaches. Between two high tides the beaches are virtually crowded by these pellets producing attractive patterns. Crabs of the genus Scopimera do not roll these sand balls at greater distances but after its formation each pellet is pushed backwards with the chelae through the legs to deposit it behind the animal (Fielder, 1970). However, the round shape of the crab’s sand pellets is just a side effect of the rotated formation by the mouthparts and it eases the transport of the pellets only secondarily.