Walking and jumping with Palaeozoic apterygote insects

[Palaeontology, Vol. 49, Part 4, 2006, pp. 827–835] WALKING AND JUMPING WITH PALAEOZOIC APTERYGOTE INSECTS by NICHOLAS J. MINTER and SIMON J. BRADDY Department of Earth Sciences, University of Bristol, Wills Memorial Building, Queen’s Road, Bristol BS8 1RJ, UK; e-mails: N.J.Minter@bris.ac.uk; S.J.Braddy@bris.ac.uk Typescript received 18 October 2004; accepted in revised form 13 June 2005 Abstract: Abundant arthropod walking and jumping traces, from the Lower Permian Robledo Mountains Formation of southern New Mexico, provide direct evidence of the locomotory techniques of monurans, an extinct group of archaeognathan apterygote insects. The jumping behaviour of monurans is compared with that of the extant machilid archaeognathan Petrobius. The jumping traces are referred to Tonganoxichnus robledoensis, and demonstrate that monurans were capable of forward progression via a linear succession of jumps of several times their body length. Petrobius also employs an unusual, fast, in-phase, jumping gait for normal directed locomotion; however, unlike the T. robledoensis traces, these jumps are only about one body length. In-phase trackways, referred to Stiaria intermedia from the Upper Carboniferous Tonganoxie Sandstone of Kansas, are found in association with Tonganoxichnus traces, indicating that monurans were also capable of such a fast jumping gait. Petrobius employs an escape jump that is more similar in terms of magnitude to those represented by T. robledoensis; however, the escape jump is essentially random in terms of direction and rotation of the body. Out-of-phase trackways from the Trace fossils are any evidence for the activity of past life, and include trackways and trails, burrows, and bite marks. They provide insights into the palaeobiology and palaeoecology of extinct animals that cannot be obtained from body fossils alone. Fossil trackways and trails provide the only direct evidence for the locomotory behaviour of extinct animals. Indeed, before the advent of high-speed cameras, traces produced by modern animals were often used to deduce their methods of locomotion (e.g. Sturm 1955; Manton 1972). Apterygote insects [archaeognathans (bristletails) and thysanurans (silverfish)] lack wings and have primitive legs with a pointed distal tarsus, as opposed to the (winged) pterygote insects in which the tarsus is placed flat on the ground (Sharov 1958; Manton 1972). The Archaeognatha are the most primitive group of true insects, body fossils of which date back to the Early Devo- ª The Palaeontological Association Robledo Mountains Formation, also referred to Stiaria intermedia, are found preceding or following on from several Tonganoxichnus traces, and demonstrate that monurans also used normal, out-of-phase, walking gaits across open ground. Analysis of these trackways demonstrates that they used a variety of gaits ranging from very slow and stable gait ratios of 1Æ2 : 8Æ8 (i.e. the propulsive backstroke phase comprises 88 per cent of the step cycle) following jumps, to fast gait ratios of 3Æ3 : 6Æ7 and 5Æ0 : 5Æ0 preceding jumps. Petrobius tends not to use such normal walking gaits unless on the undersurface of rocks, preferring to use the unusual, fast, in-phase, jumping gait over open ground. Monurans appear to have been capable of many of the same jumping behaviours as Petrobius, apart from the random escape jump. Archaeognathans are the most primitive group of true insects, and the presence of these similar types of jumping behaviours in monurans and machilids suggests that such behaviours were a primitive method of insect locomotion. Key words: Archaeognatha, locomotion, Monura, Petrobius, Tonganoxichnus, Robledo Mountains Formation. nian of Québec (Labandeira et al. 1988) and Middle Devonian of New York State (Shear et al. 1984). Rhyniognatha, from the Early Devonian Rhynie chert of Scotland, has recently been found to have affinities with pterygote insects, suggesting that true insects, including Archaeognatha, originated in the Silurian (Engel and Grimaldi 2004). The Monura are an extinct group within the Archaeognatha (Bitsch and Nel 1999; Rasnitsyn and Quicke 2002), and are known from the Upper Carboniferous–Upper Permian of Europe (Brongniart 1885; Rasnitsyn 1999), Siberia (Sharov 1957) and North America (Durden 1978), including New Mexico (Rowland 1997; Rasnitsyn et al. 2004). The reconstruction of Dasyleptus sp. from the Upper Carboniferous of Mazon Creek, Illinois (Kukalová-Peck 1987), is a composite made up of a monuran and an unspecified insect (J. Bitsch, pers. comm. 2004). Understanding the behaviour of this group 827 828 PALAEONTOLOGY, VOLUME 49 of extinct insects provides important clues to the earliest stages of insect evolution. The ichnogenus Tonganoxichnus was first described from the Upper Carboniferous Tonganoxie Sandstone of Kansas (Mángano et al. 1997), and comprises imprints of the palps, head, thoracic legs, abdominal segments and appendages, and a terminal filament. Mángano et al. (1997) erected two ichnospecies of Tonganoxichnus: T. buildexensis for resting traces and T. ottawensis for jumping traces involved with surface grazing. A third ichnospecies, T. robledoensis, was erected for locomotory jumping traces from the Lower Permian Robledo Mountains Formation of New Mexico (Braddy and Briggs 2002). Tonganoxichnus has also been reported from the Upper Carboniferous of Indiana (Mángano et al. 2001) and Oklahoma (Lucas et al. 2004). The pattern of these traces, with a head, thorax, abdomen and three pairs of thoracic appendages, indicates an insect producer; and more specifically, the imprints of pointed distal tarsi and abdominal appendages indicate an apterygote insect (Mángano et al. 1997, 2001; Braddy and Briggs 2002). The morphology of these traces is most consistent with a monuran archaeognathan (Text-fig. 1). Machilid archaeognathans possess lateral cerci, whilst thysanurans have a relatively short terminal filament with cerci perpendicular to this (Mángano et al. 1997). There is no evidence of lateral cerci in the ichnospecies of Tonganoxichnus, and the terminal filament imprint is relatively long. Rasnitsyn (1999) argued against a monuran producer because imprints of only eight abdominal segments with appendages are preserved in T. buildexensis, whilst monurans have 11 abdominal segments, nine of which have appendages. The terminal filament also appears to attach to the eighth abdominal segment. Rasnitsyn (1999) also mentioned that the imprints of the terminal filament would be more deeply impressed if they had been produced by a jumping action. However, T. buildexensis is interpreted as a resting or landing trace, and so the terminal filament may have been held above the substrate (Mángano et al. 2001). The preservation of such delicate features is likely to be variable and so the ninth abdominal segment and appendages may not have been preserved. Indeed, there is variation in the number of abdominal segments preserved between specimens of T. buildexensis from Kansas (Mángano et al. 1997) and Indiana (Mángano et al. 2001), and abdominal segments are not clear in specimens of T. robledoensis. The cylindrical imprint attached to the eighth abdominal segment may also represent an external ovipositor of a female (Mángano et al. 1997), rather than the beginnings of the terminal filament. The Robledo Mountains Formation of the Hueco Group of southern New Mexico contains one of the most abundant and diverse assemblages of Palaeozoic terrestrial T E X T - F I G . 1 . The monuran Dasyleptus (redrawn after Sharov 1957). Scale bar represents 1 mm. trace fossils in the world, including trackways of amphibians, reptiles and mammal-like reptiles, as well as various arthropod groups such as chelicerates, myriapods and insects. The formation comprises marine carbonates and shale interbedded with siliciclastic redbeds (Lucas et al. 1995, 1998), and represents a transitional zone between non-marine siliciclastic deposits to the north and marine carbonates to the south (Mack and James 1986). Conodonts indicate a late Wolfcampian (Sakmarian) age for the formation (Lucas et al. 1995, 1998). Trace fossils are most commonly preserved in red–grey siltstones to finegrained sandstones that were deposited in a broad tidalflat setting, and mudcracks and raindrop imprints indicate periods of subaerial exposure (Mack and James 1986; Hunt et al. 1993). All of the material discussed herein is housed in the New Mexico Museum of Natural History and Science, Albuquerque, New Mexico, USA (NMMNH). LOCOMOTION IN MONURANS Jumping The monuran jumping traces preserve imprints of the maxillary palps, thoracic legs, abdomen and terminal filament (Text-fig. 2), and are similar to those of the extant machilid archaeognathans Petrobius (Manton 1972; Text-fig. 3B), and Machilis (Sturm 1955; Text-fig. 4A). Braddy and Briggs (2002) recognised two types of jumping imprints, both assigned to T. robledoensis, which reflected slightly different behaviours (Text-fig. 2). Type 1 imprints show the abdomen only slightly impressed, with prominent mounds of sediment behind the thoracic leg imprints, indicating that these legs provided the main propulsive force. Type 2 imprints have more deeply impressed abdominal imprints and no obvious mounds behind the legs, indicating that the abdomen was the main propulsive organ. Jumping imprints will superimpose landing imprints, or terminate a trackway, whilst landing imprints on their own preserve detailed abdominal imprints, but lack thoracic leg imprints (Text-fig. 4B). Additional material assignable to T. robledoensis is recognised here: 23 separate sets of repeated jumping imprints, of which five preserve two jumps, one preserves three jumps and another preserves four jumps. This MINTER AND BRADDY: WALKING AND JUMPING TRACES OF PALAEOZOIC INSECTS A 829 B A, holotype and paratypes of Tonganoxichnus robledoensis (NMMNH P24020) from NMMNH locality 846, Doña Ana County, New Mexico, USA, Lower Permian, Hueco Group, Robledo Mountains Formation, preserved in negative epirelief. Succession of three Type 1 behavioural variants (including the holotype) preserved on the left of the slab, whilst a succession of three Type 2 behavioural variants are preserved on the right side. B, interpretive line drawing of T. robledoensis Types 1 and 2. Scale bar represents 10 mm. TEXT-FIG. 2. allows the jumping ability of monurans to be assessed and compared with their extant relatives. Several of these traces are preserved on mudcracked surfaces and, where observable, the mudcracks cross-cut the traces. This suggests that the traces formed whilst the substrate was still damp after exposure, and there may have been a thin film of water present. This material demonstrates that the two types of T. robledoensis of Braddy and Briggs (2002) are end-members of a morphological continuum, with some specimens having deep abdominal imprints and mounds of substrate behind the thoracic leg imprints (Textfig. 5A–B), indicating that some jumps were produced by both the abdomen and the thoracic legs. An entire morphocline of preservational variants of these imprints is present within the Robledo Mountains material, varying from complete body imprints to just those of the palps. This preservational variation may be due to undertrack fallout (Goldring and Seilacher 1971) or the animal being supported on a film of water (Mángano et al. 1997, 2001). Specimens referred to Quadrispinichna parvia Type 2 from the Robledo Mountains by Braddy and Briggs (2002) are now recognised as such preservational variants comprising imprints of the palps and the first pair of thoracic legs, and are accordingly referred to Tonganoxichnus. Numerous jumping imprints are often preserved on surfaces, and it is obviously easier to identify linear successions of jumps rather than random successions. However, repeated imprints produced by the same individual can be identified, based on their similar morphology and depth of impression (Text-fig. 5C–D). The distance of each jump was measured between the same point on consecutive imprints, and the direction of the jump was measured as the angle formed by the line between these points and the midline of the previous 830 PALAEONTOLOGY, VOLUME 49 A B A, the machilid archaeognathan Petrobius brevistylis. Petrobius varies in body length from 6 to 15 mm (redrawn after Gullan and Cranston 1994). B, escape jumping traces of Petrobius on smoked paper; the first two traces represent combined landing and jumping traces, whilst the third represents a landing trace after which the animal ran off. Scale bar represents 10 mm (redrawn after Manton 1972). TEXT-FIG. 3. A B T E X T - F I G . 4 . A, jumping trace of the machilid archaeognathan Machilis. B, landing trace of Machilis (both redrawn after Sturm 1955). Note the clear imprints of the thoracic appendages and maxillary palps in the jumping trace whereas the imprints of the abdominal segments are more clearly visible in the landing trace. Scale bars represent 10 mm. imprint. Rotation of the body was measured as the angle between the midlines of the imprints. The majority of jumps demonstrate a forward progression, most 50– 80 mm apart, but some more than 120 mm, and in a sector 45 degrees to either side (Text-fig. 6). Most jumps also show only a small degree of rotation of the midline of the body from its original orientation. The body length of archaeognathans is typically measured from the head to the back of the abdomen, excluding the palps and terminal filament. Head imprints are not preserved in the Robledo Mountains material; however, comparisons with the Upper Carboniferous traces from Kansas and Oklahoma, in which the head imprints are present, suggests comparable body lengths of about 15–20 mm, slightly larger than reported monuran body fossils (Sharov 1957; Bitsch and Nel 1999; Rasnitsyn 1999), and extant machilids (Delany 1954). Most jump distances were therefore 3–5 times body length, but can be as much as 6–8 times. Specimens showing successions of two or more consecutive jumps are predominantly in a forward direction (Text-figs 2, 5C–D), as confirmed by the orientation of the imprints and the presence of sediment mounds behind the thoracic leg imprints of some specimens. This suggests that monurans were capable of forward progression via a succession of jumps of several times the body length. When disturbed, for example, by a predator, Petrobius performs a bewildering succession of jumps of random distance and direction, during which the body may rotate in the air, and a statistically significant number of landings occur after body rotation of about 160–180 degrees (Evans 1975). These jumps also tend to be around 50 mm apart, but can be as much as 110 mm. They are produced by an initial depression of the posterior of the abdomen and the terminal filament, which raises the centre of gravity, followed by ventral flexure of the thorax, the rapidity of which determines whether the jump is forwards or backwards (Evans 1975). These jumps are used as a predator escape mechanism by Petrobius, the random nature of which makes it difficult for the predator to predict their movements. The Type 2 end-member monuran jumping imprints appear to have been produced by the abdomen, as in the escape jump of Petrobius; however, these jumps are in an overall forwards direction, and the degree of rotation is lower than in the escape jumps. These jumps are at the lower end of the range of distances, suggesting that there may have been more of a vertical component to the jump than with those produced by the thoracic appendages. There is no evidence in the Robledo Mountains Formation of random escape jumps. Petrobius also performs an unusual fast jumping gait when moving across open ground, whereby the paired legs move in-phase, and the jump distance is typically one body length (Manton 1972). The Type 1 end-member MINTER AND BRADDY: WALKING AND JUMPING TRACES OF PALAEOZOIC INSECTS A B C 831 D A, specimen NMMNH P24018 from NMMNH locality 846, Doña Ana County, New Mexico, USA, Lower Permian, Hueco Group, Robledo Mountains Formation, demonstrating intergrading between Type 1 and Type 2 behavioural variants, with deeply impressed abdominal imprints and mounds of substrate behind the thoracic leg imprints. B, interpretive line drawing of specimen NMMNH P24018. C, specimen NMMNH P24242 from NMMNH locality 2821, Doña Ana County, Robledo Mountains Formation, preserving two separate successions of jumping imprints; note that it is possible to identify particular successions based on the morphology of the traces. D, interpretive line drawing of specimen NMMNH P24242. Both specimens are preserved in negative epirelief. Scale bars represent 10 mm. TEXT-FIG. 5. imprints appear to have been produced by the thoracic legs, but unlike the jumping gait of Petrobius, they are several body lengths in terms of distance. Trackways referred to Stiaria intermedia from the Upper Carboniferous of Kansas (Buatois et al. 1998) occur on the same surfaces as Tonganoxichnus traces. These trackways comprise series of 2–3 imprints and are in-phase. These trackways are very similar to modern traces produced by the fast jumping gaits of Petrobius (Manton 1972; Text-fig. 7), and Buatois et al. (1998) suggested that they could have been produced by a fast, in-phase, jumping gait of a monuran. Walking Trackways provide the only direct evidence of how extinct animals walked, and assuming the producer and its stance are known, the animal’s walking techniques can be reliably reconstructed. Studies of arthropod locomotion, including hexapods, were pioneered by Manton (1952, 1972). The step cycle can be divided into a backward, propulsive, retraction period (r) and a forward, recovery, protraction period (p). The regulation of arthropod leg movements is governed by three parameters: the gait ratio (p : r), usually expressed as a proportion of ten; the opposite phase difference (opp; the proportion of the step cycle that right legs move after the left legs); and the successive phase difference (suc; the proportion of the step cycle that a leg moves before the leg in front) (Manton 1952; Braddy 2001). Hexapods, with only three pairs of legs, are more limited than, for example, myriapods, in the gait patterns that they can employ because at least three legs must be in contact with the ground at any one time in order to 832 PALAEONTOLOGY, VOLUME 49 T E X T - F I G . 6 . Radar plot of the distance and direction of jumps, as measured to the midpoint of the arrow. The orientation of arrows from the 0
line indicates the degree of rotation of the midline of the body from its original position. Specimens: NMMNH P23598 (two separate successions), P23602, P23638, P23897, P23915 (two separate successions), P24018, P24020 (three separate successions), P24059, P24062, P24212, P24242 (two separate successions), P24319, P24323, P24328, P24348–50 and P24485. A B T E X T - F I G . 7 . A–B, in-phase trackways of the fast jumping gait of P. brevistylis on smoked paper (redrawn after Manton 1972). Scale bars represent 5 mm. remain stable (Manton 1972). This means that the propulsive backstroke cannot be shorter than the forward recovery stroke, otherwise there will be periods of instability with only two legs in contact with the ground. Five trackway specimens from the Robledo Mountains Formation can be confidently attributed to monurans, because they are terminated by jumping imprints or follow landing imprints (Text-figs 8–9), although two of these are too poorly preserved to allow analysis. These trackways have external widths of 7Æ2–11Æ3 mm, and track series that are out-of-phase. The track series generally diverge backwards at an angle of 45 degrees but occasionally diverge forwards or are orientated perpendicular to the trackway axis. The rounded tracks further support that they were produced by apterygote insects, which had pointed distal tarsi. The curvilinear marks situated anterior to the tracks were produced by the distal tip of the leg scraping the substrate during its recovery stroke. These trackways are referred to Stiaria intermedia. S. intermedia is similar to Paleohelcura tridactyla in that they comprise series of 2–3 imprints with a medial impression, although the medial impression is not always present in P. tridactyla. Stiaria also tends to be narrower than Paleohelcura, and the external widths of these specimens are more similar to those of S. intermedia. The track series of S. intermedia tend to be orientated perpendicular to the trackway axis or diverge backwards, as in these specimens, whereas those of P. tridactyla diverge forwards, although it is often the case that the diverging direction cannot be ascertained because there are no indicators of the direction of travel. The walking techniques of the monurans that produced these fossil trackways may be assessed by calculating the gait parameters, based on trackway data in conjunction with a body plan of the producer, according to the following formulae: r ¼ ðbackstroke=strideÞ
10; p ¼ 10  r; opp: ¼ pace=stride (Braddy 2001): The stride, as measured from a trackway, is the distance between successive imprints of the same leg, whilst the pace is the longitudinal distance between the imprints of opposing legs. The backstroke distance of Manton’s (1972) body plan for Petrobius was used for the monuran backstroke distance by equating the ‘straddle’ (i.e. the maximum transverse distance between the outer pair of legs) of Petrobius to the external width of the trackways. Petrobius was used because it is the closest extant equivalent for a monuran, and it has a similar leg structure and stance to monurans (Sharov 1966). The successive phase difference varied, as evident from the orientation of the track series, which is governed by the gait ratio and the successive phase difference although the gait ratio is likely to have been constant because the stride is consistent within a trackway. One of the best preserved trackways (Text-fig. 8) follows a landing trace and shows the slowest gait (i.e. p : r of 1Æ2 : 8Æ8, opp. 0Æ27), meaning that each leg spends 88 per cent of the step cycle in the propulsive backstroke phase. This would have been a very stable gait. One specimen preceding a jumping imprint shows the fastest gait (i.e. p : r 5Æ0 : 5Æ0, opp. 0Æ4), the maximum for a hexapod if stability is to be maintained, whilst another (Textfig. 9) has an intermediate gait ratio (i.e. 3Æ3 : 6Æ7, opp. MINTER AND BRADDY: WALKING AND JUMPING TRACES OF PALAEOZOIC INSECTS A B 833 0Æ3). Trackways of similar morphology are relatively common in the Robledo Mountains Formation, suggesting that normal, out-of-phase, walking was a common A B A, Stiaria intermedia, walking trackway of a monuran, specimen NMMNH P24292 from NMMNH 2851, Doña Ana County, New Mexico, USA, Lower Permian, Hueco Group, Robledo Mountains Formation, preserving a negative epirelief trackway terminated by a jumping imprint (stride, 7Æ3 mm; pace, 2Æ2 mm; external width 9Æ8 mm; backstroke, 4Æ9 mm). Note the variation in form of the series of imprints from linear series diverging forwards at the beginning of the preserved length of the trackway to transverse series and series diverging backwards. B, interpretive line drawing of specimen NMMNH P24292. Scale bar represents 10 mm. TEXT-FIG. 9. Stiaria intermedia, walking trackway of a monuran, specimen NMMNH P24019 from NMMNH locality 846, Doña Ana County, New Mexico, USA, Lower Permian, Hueco Group, Robledo Mountains Formation, preserving a short negative epirelief trackway terminated by a jumping imprint, followed by a landing imprint which precedes a trackway (stride, 6Æ4 mm; pace, 1Æ7 mm; external width, 11Æ3 mm; backstroke, 5Æ65 mm). Note the well-defined abdominal imprint and lack of thoracic leg imprints in the landing imprint. B, interpretive line drawing of specimen NMMNH P24019. Scale bar represents 10 mm. TEXT-FIG. 8. 834 PALAEONTOLOGY, VOLUME 49 method of monuran locomotion, although they lack associated jumping and landing imprints that would enable them to be confidently attributed to monurans. Petrobius also employs out-of-phase walking gaits, but only when on the undersurfaces of rocks where it cannot use the fast, in-phase, jumping gait, and gait ratios between 2Æ5 : 7Æ5 and 4Æ0 : 6Æ0 have been recorded (Manton 1972). CONCLUSIONS Trace fossils provide the only direct evidence for the locomotion of extinct animals. Abundant arthropod walking and jumping traces in the Lower Permian Robledo Mountains Formation of southern New Mexico are attributable to monurans, an extinct group of apterygote archaeognathan insects. These traces allow their range of walking gaits to be deduced and indicate that they were capable of forward progression via a succession of jumps of several times their body length. A morphological continuum of jumping imprints with two end-members is recognised. The Type 1 end-member imprints have shallow abdominal imprints and mounds of substrate behind the thoracic leg imprints, indicating that the thrust was provided by the legs, whereas the Type 2 imprints have deeper abdominal imprints indicating that they were produced by depression of the abdomen. The extant machilid archaeognathan Petrobius employs a fast jumping gait for normal locomotion in which the paired thoracic legs move in-phase to produce the thrust as in the Type 1 monuran jumps; however, they are typically only of one body length. Trackways referred to Stiaria intermedia from the Upper Carboniferous of Kansas occur on the same surfaces as Tonganoxichnus traces, and are in-phase, suggesting that monurans were also capable of a fast, in-phase, jumping gait similar to that of Petrobius (Buatois et al. 1998). Petrobius also has an escape jump produced by the abdomen, as in the Type 2 monuran jumps; however, these escape jumps are essentially random, and some are backwards, with significant rotation of the body. The random nature of these jumps makes it difficult for predators to predict their movements. Fossil trackways from the Robledo Mountains Formation demonstrate that monurans commonly walked across open ground using normal out-of-phase gaits. Trackways that precede jumps show intermediate to fast gait ratios of 3Æ3 : 6Æ7 and 5Æ0 : 5Æ0, whereas one that follows a jump demonstrates a slower gait ratio of 1Æ2 : 8Æ8. Most modern jumping insects jump from a static position after storing up energy for the jump. There is no shortening of stride length in the trackways preceding jumps (Text-figs 8–9) as would be expected if the animal was slowing, and so monurans appear to have literally leapt into them. Monurans appear to have been capable of many of the same jumping behaviours of Petrobius, apart from the random escape jump. Petrobius appears to have largely dispensed with an out-of-phase walking gait, only using it on the undersurfaces of rocks. Only five trackways from the Robledo Mountains Formation can be definitively attributed to monurans because they have landing or jumping traces preserved at their ends, although there are many other trackways of similar morphology, suggesting that normal walking was more common in monurans than in machilids. No evidence of random escape jumps has been found so far, although other traces in the Robledo Mountains Formation, such as reptile, amphibian, mammal-like reptile and chelicerate trackways, provide evidence for many potential predators. Monurans may, therefore, have tended to flee predators by these linear jumps, a behaviour present in more primitive arthropod groups such as millipedes (Evans and Blower 1973). Evidence from trackways indicates that monurans used intermediate to fast gaits prior to a jump, probably to build up momentum before launching into a jump, and such jumps appear to have been produced by the thoracic legs. Conversely, trackways following jumps demonstrate much slower and more methodical walking techniques. Such linear fleeing methods may therefore have been a more primitive behaviour than random escape jumps. Archaeognathans are the most primitive group of true insects, and the presence of these similar types of jumping behaviours in monurans and machilids suggests that such behaviours were a primitive method of insect locomotion. Acknowledgements. We thank Dr S. G. Lucas, Dr A. B. Heckert, Dr A. P. Hunt, Dr G. S. Morgan and A. J. Lerner (NMMNH) for access to specimens and discussion of material. We also thank Dr M. G. Mángano and Dr L. A. Buatois for discussions on the material, and Prof. J. Bitsch for discussion on monuran affinities. We are grateful to Dr N. H. Trewin, Dr A. P. Rasnitsyn and Dr M. G. Mángano for reviewing the manuscript. We also thank E. K. Sumner for translating the Russian literature. NJM is funded by NERC studentship NER ⁄ S ⁄ A ⁄ 2003 ⁄ 11199. SJB acknowledges a Palaeontological Association Sylvester-Bradley Award, which enabled initial investigation of the material. REFERENCES B I T S C H , J. and N E L , A. 1999. Morphology and classification of the extinct Archaeognatha and related taxa (Hexapoda). Annales de la Société Entomologique de France, 35, 17–29. B R A D D Y , S. J. 2001. Trackways – Arthropod locomotion. 389–393. In B R I G G S , D. E. G. and C R O W T H E R , P. R. (eds). Palaeobiology II. Blackwell, Oxford, 583 pp. —— and B R I G G S , D. E. G. 2002. New Lower Permian nonmarine arthropod trace fossils from New Mexico and South Africa. Journal of Paleontology, 76, 546–557. MINTER AND BRADDY: WALKING AND JUMPING TRACES OF PALAEOZOIC INSECTS B R O N G N I A R T , C. H. 1885. Les insectes fossiles des terrains primaries. Bulletin de la Société des Amis des Sciences Naturelles de Rouen, 1885, 50–68. B U A T O I S , L. A., M Á N G A N O , M. G., M A P L E S , C. G. and L A N I E R , W. P. 1998. Ichnology of an Upper Carboniferous fluvio-estuarine paleovalley: the Tonganoxie Sandstone, Buildex Quarry, eastern Kansas, USA. Journal of Paleontology, 72, 152– 180. D E L A N Y , M. J. 1954. Studies on the life history and ecology of Dilta littoralis (Womersley, 1930) (Thysanura, Machilidae). Transactions of the Royal Entomological Society of London, 105, 31–63. D U R D E N , C. J. 1978. A dasyleptid from the Permian of Kansas, Lepidodasypus sharovi, n. gen., n. sp. (Insecta: Thysanura: Monura). Pearce-Sellard Series, 30, 1–9. E N G E L , M. S. and G R I M A L D I , D. A. 2004. New light shed on the oldest insect. Nature, 427, 627–630. E V A N S , M. E. G. 1975. The jump of Petrobius (Thysanura, Machilidae). Journal of Zoology, 176, 49–65. —— and B L O W E R , J. G. 1973. A jumping millipede. Nature, 246, 427–428. G O L D R I N G , R. and S E I L A C H E R , A. 1971. Limulid undertracks and their sedimentological implications. Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen, 137, 422–442. G U L L A N , P. J. and C R A N S T O N , P. S. 1994. The insects: an outline of entomology. Chapman & Hall, London, 491 pp. H U N T , A. P., L O C K L E Y , M. G., L U C A S , S. G., M A C D O N A L D , J. P., H O T T O N , N. III and K R A M E R , J. 1993. Early Permian tracksites in the Robledo Mountains, south-central New Mexico. New Mexico Museum of Natural History and Science, Bulletin, 2, 23–31. K U K A L O V Á - P E C K , J. 1987. New Carboniferous Diplura, Monura, and Thysanura, the hexapod ground plan, and the role of thoracic side lobes in the origin of wings (Insecta). Canadian Journal of Zoology, 65, 2327–2345. L A B A N D E I R A , C. C., B E A L L , B. S. and H U E B E R , F. M. 1988. Early insect diversification: evidence from a Lower Devonian bristletail from Québec. Science, 242, 913–916. L U C A S , S. G., A N D E R S O N , O. J., H E C K E R T , A. B. and H U N T , A. P. 1995. Geology of Early Permian tracksites, Robledo Mountains, south-central New Mexico. New Mexico Museum of Natural History and Science, Bulletin, 6, 13–32. —— H E C K E R T , A. B., E S T E P , J. W., H U N T , A. P. and A N D E R S O N , O. J. 1998. Stratigraphy, paleontology and depositional environments of the Lower Permian Robledo Mountains Formation of the Hueco Group, Robledo Mountains, New Mexico. New Mexico Museum of Natural History and Science, Bulletin, 12, 29–42. —— L E R N E R , A. J., B R U N E R , M. and S H I P M A N , P. 2004. Middle Pennsylvanian ichnofauna from eastern Oklahoma, USA. Ichnos, 11, 45–55. 835 M A C K , G. H. and J A M E S , W. C. 1986. Cyclic sedimentation in the mixed siliciclastic–carbonate Abo-Hueco transitional zone (Lower Permian), southwestern New Mexico. Journal of Sedimentary Petrology, 56, 635–647. M Á N G A N O , M. G., B U A T O I S , L. A., M A P L E S , C. G. and L A N I E R , W. P. 1997. Tonganoxichnus, a new insect trace from the Upper Carboniferous of eastern Kansas. Lethaia, 30, 113–125. —— L A B A N D E I R A , C. C., K V A L E , E. P. and B U A T O I S , L. A. 2001. The insect trace fossil Tonganoxichnus from the Middle Pennsylvanian of Indiana: paleobiologic and paleoenvironmental implications. Ichnos, 8, 165–175. M A N T O N , S. M. 1952. The evolution of arthropodan locomotory mechanisms: Part 2. General introduction to the locomotory mechanisms of the Arthropoda. Zoological Journal of the Linnean Society, 42, 93–117. —— 1972. The evolution of arthropodan locomotory mechanisms: Part 10. Locomotory habits, morphology and evolution of hexapod classes. Zoological Journal of the Linnean Society, 51, 203–400. R A S N I T S Y N , A. P. 1999. Taxonomy and morphology of Dasyleptus Brongniart, 1885, with description of a new species (Insecta: Machilida: Dasyleptidae). Russian Entomological Journal, 8, 145–154. —— and Q U I C K E , D. L. J. 2002. History of insects. Kluwer Academic Publishers, Dordrecht, 517 pp. —— A R I S T O V , D. S., G O R O C H O V , A. V., R O W L A N D , J. M. and S I N I T S H E N K O V A , N. D. 2004. Important new insect fossils from Carrizo Arroyo and the Permo-Carboniferous faunal boundary. New Mexico Museum of Natural History and Science, Bulletin, 25, 215–246. R O W L A N D , J. M. 1997. The late Paleozoic insect assemblage at Carrizo Arroyo, New Mexico. New Mexico Museum of Natural History and Science, Bulletin, 11, 1–7. S H A R O V , A. G. 1957. Peculiar Palaeozoic wingless insects belonging to a new order Monura (Insecta, Apterygota). Doklady Akademii Nauk SSSR, 115, 795–799. [In Russian]. —— 1958. The structure of extremities and mode of locomotion in Monura and Thysanura (Insecta, Apterygota). Doklady Akademii Nauk SSSR, 122, 733–736. [In Russian]. —— 1966. Basic arthropodan stock. Pergamon Press, Oxford, 272 pp. S H E A R , W. A., B O N A M O , P. M., G R I E R S O N , J. D., R O L F E , W. D. I., S M I T H , E. L. and N O R T O N , R. A. 1984. Early land animals in North America: evidence from Devonian age arthropods from Gilboa, New York. Science, 224, 492–494. S T U R M , H. 1955. Beiträge zur ethologie einiger mitteldeutscher Machiliden. Zeitschrift für Tierpsychologie, 12, 337–363.