Restoring the Kalligrammatids: The Not-Butterflies of the Mesozoic

 

Fig. 1 Representative kalligrammatids showing the diversity of the family.

Recently I’ve been working on a set of illustrations featuring the kalligrammatids, a family of Mesozoic lacewings which are most well known for their butterfly-like appearance and wing eyespots, and while I’ve been doing it I’ve realised that there are no good references or guides to their life appearance available to palaeoartists and people just interested in these insects. So, while this is by no means a comprehensive guide, here are a few quick notes on what I’ve learnt about the kalligrammatids and how to restore them. I’d like to start out by saying that I am not an invertebrate specialist; as some of you will know, and many of you will be able to guess from the name of my blog, I am a palaeobotanist, so my specialist knowledge is on plants, not insects, so if any of you more adept in the ways of the bug notice anything glaringly wrong here please let me know!

            It is a common trend within palaeoart that certain groups get WAY more attention than others (I’m looking at you, non-avian dinosaurs), and any group which is even the tiniest bit obscure often has no easily accessible information about it. Kalligrammatidae is one of these groups. Yet despite this they are reasonably common elements within palaeoart (albeit often merely as background elements in a scene featuring something larger and toothier). This is unsurprising, as they are some of the showiest insects of the Mesozoic, and a number of well-known studies with a lot of news coverage have been released in the last ten years or so (Labandeira et al., 2016; Yang et al., 2014). However, despite their relatively common inclusion within Mesozoic scenes, very few people have explored the diversity of kalligrammatids which were present. An unfortunate trope when restoring kalligrammatids is to make them almost butterfly-clones; this is not that surprising, as they have been ‘marketed’ as being very butterfly-like, and indeed they are (I was actually a little surprised by how superficially similar some taxa are to modern lepidopterans)! However, this similarity is only very superficial, and closer examinations of their fossils and other neuropterans (the net-winged insects, the group to which the kalligrammatids belonged) reveals that their anatomy was rather different from butterflies.

 

So, what are the kalligrammatids?

Well, first off, let me introduce you to the kalligrammatids: the kalligrammatids are an extinct family of winged insects within the order Neuroptera, a clade which in the modern world includes things like lacewings, mantidflies (Mantispidae) and antlions (Myrmeleontidae). They originated in the late-Early Jurassic (~175-180 Ma) (Ansorge and Markarkin, 2020) and went extinct in the mid-Early Cretaceous (~124 Ma) (Bechly and Makarkin, 2016; Yang et al., 2014; Zhong et al., 2021). Almost all species originate from Eurasia, although three species are also known from the Crato Formation of northern Brazil, suggesting their distribution was larger than is currently known (Machado et al., 2021). Several small neuropterans have been found in the Myanmar amber, however their taxonomy is uncertain; these insects were initially considered to form their own clade, outside of Kalligrammatidae, but within the same superfamily (Psychopsoidea, of which the only extant family is Psychopsidae, the silky lacewings) (Lu et al., 2016), however, a subsequent study found that these, and several new genera, formed a small subfamily (Cretanallachiinae) within Kalligrammatidae (Liu et al., 2018). If these small neuropterans do turn out to be kalligrammatids, this would extend the known geological and chronological range of the family, however, because of their taxonomic uncertainty (and slightly different appearance) I will not be discussing them other than to point out the following:

·         They are much smaller than other known kalligrammatids; almost 1/10^th the size of even the smallest kalligrammatids I’ve illustrated here.

·         The males have bipectinate (frilly, comb-like) antennae (which aren’t known from other kalligrammatids; Yang et al., 2014).

·         They have elongated proboscis-like mouthparts so were nectar feeders.

·         They have two pairs of palps (feeding appendages), maxillary and labial, unlike other kalligrammatids which appear to only have one pair (Labandeira et al., 2016 supp. Info.).

·         And some species had eyespots on their wings (e.g. Burmogramma liui), although most have clear wings (Liu et al., 2018).

 

Kalligrammatid anatomy

When doing my recent illustrations, I started out by producing an anatomical diagram of one species, so that I could get to grips with their anatomy and life appearance (Fig. 2). This was interesting to do, as I’m not aware of this having been done before, so my restoration is based mainly on the fossils we have (which often preserve the wings in a lot of detail, but with disarticulated bodies), with other details supplemented by looking at modern neuropterans. Some of the main points which I found important when restoring kalligrammatids are as follows:

·         Their thorax (the middle section of their body) is split into three sections (the pro or protothorax, the mesothorax, and the metathorax) and the forewings attach to the mesothorax, and hindwings to the metathorax.

·         The two pairs of wings often overlap and the forewings overlap the hindwings dorsally (when viewed from above the forewings are on-top).

·         The bases of their wings are cordate where they attach to the body, meaning that they’re heart-shaped, with two bulges, in-front of and behind the point of attachment – forgive me for using a botanical term (‘cordate’), there may be a formal term for this in insect wings, but I don’t know it.

·         Their eyes were often large, sometimes almost meeting in the middle of the head.

·         Their wings were covered in microscopic scales, which would have given the wings colour; however…

·         Their wings were likely most often translucent, like most modern neuropterans, as most species do not seem to have a thick covering of scales; however…

·         Some species DID have a thick covering of scales, which in the fossils completely, or near-completely obscure the venation, and these taxa likely had opaque wings (e.g. Abrigramma calophleba; Sophogramma lii, Ithigramma spp. and Meioneurites spectabilis).

·         Their antennae are filiform (thread-like and segmented) and often long, though not as long as the forewings, and short in some genera (e.g. Ithigramma).

·         Their bodies likely had a sparce covering of hairs, although these may have been localised and some diversity in the distribution, density and length of these hairs would be expected (I have not observed long hairs in any fossils, so their bodies would not have been fluffy like moths; you can look at modern neuropterans for inspiration).

·         Their legs may also have had a sparce covering of hairs.

·         They probably didn’t have long tibial spurs, although short ones may have been present in some species (e.g. they’re known from the fore legs of Meioneurites spectabilis). As a side note, this was something I only noticed while writing this, after having finished all the diagrams, so I had to go and erase them from all my drawings (apologies if I missed some).

·         Long, curved ovipositors are known from at least one species (Oregramma illecebrosa), but they seem to be absent in most taxa, even when the specimens are presumably female and well preserved. This means that kalligrammatids should not be ubiquitously restored with long ovipositors.

·         The length of their abdomens varies, but none extend beyond the length of the hindwings, and the abdomens of males were likely very short (see Oregramma sp.; Yang et al., 2014).

·         Most kalligrammatids had a proboscis (siphonate mouthparts), with mandibulate mouthparts being restricted to basal clades (Sophogrammatinae).

 

Fig. 2 Anatomy of a kalligrammatid (in this case Oregramma illecebrosa).

The diversity of nectar-feeding kalligrammatids

The thing which started my recent obsession with kalligrammatids was when I began doing an illustration showing the diversity of species which would have fed on nectar and been plant pollinators (Fig. 3). All of the species I have restored are known to have had a proboscis, or siphonate mouthparts (Yang et al., 2014) with the exception of Markarkinia, which is only known from wing fragments. I included two species from this genus to show that some kalligrammatids could get very large; the largest are estimated to have had a wingspan of ~30 cm across, with individual wings being 16 cm long (Machado et al., 2020), however here I have been a little more conservative, giving them only ~25 cm wingspans. This is something which I was consistently surprised by when restoring these kalligrammatids: exactly how large most of them where!!! Of the species which I looked at in detail, the smallest ones still had wingspans of about 9 cm, the size of the UK’s largest butterfly, the swallowtail, and most had wingspans in the range of 12 to 19 cm across, much larger than any modern neuropterans, which have wingspans maxing out around 10 cm across. These large sizes are quite alarming for insects and is something worth considering when restoring kalligrammatids. It is also worth thinking about when pairing kalligrammatids with plants; in fact, it may help explain why some bennettitalean ‘floral organs’ were so large (e.g. Weltrichia sol had a ‘floral organ’ up to 20 cm across; Popa, 2019).

Fig. 3 Diversity of nectar-feeding kalligrammatids. a. and b. from the Crato Formation (Brazil); c., d., e., f., g., h., p. from the Jiulongshan or Haifanggou Formation (China); i., j., k., l., m., n. from the Yixian Formation (China); o. from the Karabastau Formation (Kazakhstan).

            Something else I found fascinating when restoring the kalligrammatids was just how morphologically diverse they actually were. I am so used to seeing the one or two different restorations being used over and over again that I was surprised at how much their appearance varied from species to species. The most obviously variable thing about them was their wing shapes (see Fig. 3); as a general rule, the forewings are always longer than the hindwings, and often narrower, with a higher aspect ratio, and the hindwings were much broader and often rounded or almost triangular-shaped. However, some species (e.g. Meioneurites spectabilis and Affinigramma myrioneura) had forewings which were incredibly broad and almost completely hid the smaller hindwings, creating an almost skirt-like silhouette when the wings were spread out. Still other species, such as those within Kallihemerobius possessed hind and forewings which were almost identical in size and shape. In Oregramma and Abrigramma the forewings are very long and narrow with pointed tips, and their hindwings are triangular-shaped and large.

            One of the most unexpectedly variable features within the kalligrammatids was the length and size of their prothorax (also called a protothorax; Fig. 2). The prothorax is the first segment of their thorax, and in some modern neuropterans serves almost as a neck, which can be moved up and down relative to the rest of the body (the forelegs also attach to this body section). In all kalligrammatids the top of the prothorax appears to be almost shield-like, with a smooth curved surface (actually, now that I think about it, it’s more saddle-like) which extends down its sides. In most species it is pretty innocuous, being similar to that seen in Oregramma illecebrosa in Fig. 2; however, in some species it is incredibly long and robust, causing the head to project out in front of the wings! This is most obvious in Ithigramma sp. (Yang et al., 2014), although it can also be observed to a lesser extent in Oregramma sp. and Abrigramm calophleba. Meioneurites spectabilis also has a longer, but also narrower, prothorax. Variability in the length of the prothorax is a feature also commonly seen in modern neuropterans, with extreme examples being seen in the mantidflies and snakeflies, so it’s interesting that the prothorax of kalligrammatids is also very variable.

            The shape and size of kalligrammatid abdomens does vary somewhat, although most are short and broad. Some genera did possess longer, narrower abdomens, particularly Kallihemerobius (which had long abdomens which were broadest near their base and narrowed gradually towards their end) and Meioneurites (which had an abdomen which almost extended beyond the back-edge of the wings and was an elongate oval shape with a blunt end; Engel, 2005). Some kalligrammatids possessed long, curved ovipositors, as seen in Oregramma illecebrosa, but the distribution of these within the group is not known, and it likely was not a ubiquitous feature, as multiple taxa with well preserved abdomens show no sign of an ovipositor (although the presence/absence of ovipositors is also confused by the fact that male specimens are difficult to identify as male unless the preservation is superb) (Engel, 2005; Yang et al., 2014). As a general rule I would say, if in doubt, don’t include an ovipositor when restoring kalligrammatids.

            The head shape of most kalligrammatids is pretty consistent, although details such as the length of the proboscis and palps, antenna length and overall head size do vary somewhat. These variations can be seen in Fig. 3. (although I appreciate at this scale it’s a little difficult to see). As mentioned earlier, several taxa of kalligrammatids (e.g. Sophogramma lii) did have mandibulate mouthparts, however I have not looked at these taxa in detail, so cannot talk about their appearance.

            Leg length was also incredibly variable within the kalligrammatids, although this cannot be discussed in much detail as few specimens preserve legs in any great detail. Most taxa with preserved legs seem to have reasonably short ones (e.g. Ithigramma and Abrigramma) which ended in two curved claws and (seemingly) lacked large tibial spurs (Yang et al., 2014). The legs of Meioneurites spectabilis, while also short, were very robust, and possessed short tibial spurs on at least the first two pairs of legs; it also possessed large claws at the end of its short robust tarsus. Another species worth discussing is Kallihemerobius almacellus; while the legs of this species are not very well preserved, they indicate that this species had very long legs (Yang et al., 2014).

            Perhaps the most beautiful feature of at least the fossils of kalligrammatids is that most preserve at least some patterning! These patterns most commonly (and most famously) take the form of eyespots, which are present within almost all kalligrammatids. Because of their wide distribution, it is unsurprising that the forms of these eyespots vary quite significantly between the different genera, and even between species of the same genera. Five different forms of eyespots (named Type 1 - 5) were described by Labandeira et al. (2016) and these range from simple black spots, as seen in Ithigramma and some species of Kallihemerobius, all the way to complex ‘eyes’ which incorporate eyeshine-spots and multiple rings, like those in Oregramma illecebrosa, Makarkinia, Kalligramma, and some species of Kallihemerobius. One of the most fascinating wings I came across was that of Kalligramma brachrhyncha. When restoring this species I noticed that the holotype preserves a series of fain pale slanting-triangles along the anterior edge of the wing, which in the fossil appear pretty innocuous; however, when adding these pale patterns to my restoration (Fig. 3h, 5) I was surprised when I was greeted by a toothy grin! While a closer examination of the fossil would be needed to confirm that this pattern is genuine, this could represent a remarkable level of mimesis, with the wings mimicking the head of a small predator (small dinosaur anyone?) when folded closed. Something else which I found interesting when restoring multiple genera is that similar genera often share similar types or distributions of eyespots. This is unsurprising, but still interesting to see such consistent patterns within fossil genera; this consistency is useful when restoring incomplete species from the same genera. For instance, the fore and hindwings of Kallihemerobius aciedentatus are known and they seem to indicate that both fore and hindwings had eyespots (although Labandeira et al., 2016 interpreted the fossil as having two spots on the forewing only); in the other two species of Kallihemerobius which I restored (K. almacellus and K. feroculus) only either the fore or hindwings are known from each species, however both possess complex eyespots so it is reasonable to assume that eyespots would have been present on both fore and hindwings, like in Kallihemerobius aciedentatus. In Ithigramma species, the eyespots are simple black spots and appear to be present on both the fore and hindwings of both species (although they are obscured by the forewings in Fig. 3j). The eyespots in Kalligramma species are quite variable, but consistently have a central black spot with smaller eyeshine-spots which are surrounded by a pale region enclosed within a dark ring. The eyespots of Kalligramma circularia are difficult to interpret from the fossil, but this species either has two eyespots per wing (as I have shown here) or they had a single eyespot on the forewing, and a smaller one on the hindwing. Both Makarkinia species which I have shown are known from the hindwings, and both possess large eyespots with a simple black centre encircled by three dark rings of varying thickness. Finally, in Oregramma only one species seems to preserve an eyespot, however the hindwings of both Oregramma illecebrosa and Oregramma aureolusa seem to have light and dark barring. As previously mentioned, the wings of both Abrigramma and Meioneurites were very densely covered in scales, so it is likely that in life the wings of these species were opaque, possibly with very rich colouration. Simple dark eyespots are apparently present in Abrigramma calophleba (Labandeira et al., 2016), however I could not identify them or their position in the fossil. Affinigramma was the only genus I restored for which I could find no evidence of eyespots.

 

Kalligrammatid flight (probably wasn’t butterfly-like)

As previously mentioned, the kalligrammatids were neuropterids, so the following section is mainly based on observations of modern neuropterid flight, and then applying that to the kalligrammatids. When thinking about kalligrammatid flight I looked mainly at videos on youtube showing the take-off and flight of a variety of modern neuropterids. Of particular use were videos by the users NickSchrader and Ant Lab (video one and two). I’ve outlined my conclusions below (and in a rather janky animation I made; Fig. 4), but it is also worth pointing out that this is my own interpretation, and I’m not a biomechanist by any means:

·         When taking flight the kalligrammatids would start by unfolding their wings and spreading them out.

·         They would then raise their wings up while also tilting up their prothorax, lifting their front pair of legs off the ground (this is something which all modern neuropterans apparently do, as it moves the legs out of the way of the wings).

·         The kalligrammatids would then flap down, moving their forewings down, followed by their hindwings a moment later.

·         This downstroke would gently lift them off the surface and they would continue to flap to gain speed and elevation.

·         During the upstroke the fore and hindwings would separate, with the forewings being lifted up more quickly than the hindwings.

·         Flapping frequency would probably be very low, as in larger butterflies, probably within the range of 9-12 beats per second (Zhang et al., 2021).

·         All four wings could be moved independently and in normal flight the flapping cycles of the hindwings would have lagged behind the forewings.

·         During the downstrokes the forewings would sweep down and slightly forward under the body, and the hindwings would sweep down under the body, almost touching the opposite wing.

·         Like butterflies, they may have periodically glided, especially in the larger species, but smaller species more likely employed continuous flapping like modern neuropterans.

·         Unlike most modern neuropterans, kalligrammatids with wings which overlap for most of their surface probably kept their wings together on the downstroke.

·         I imagine that they were probably quite awkward when landing, due to their large size and relatively small legs, but this is more speculative and can’t really be tested, It’s more of just a personal hunch.

 Fig. 4 An animation showing the take-off sequence for a kalligrammatid in slow motion. Note the unfolding of the wings from over the body, the lifting up of the front legs before takeoff, and the slight delay in the flap cycle of the hindwings.

Despite their visual similarities to butterflies, and despite often being illustrated as almost butterfly-clones, their flight would likely have appeared quite different. Butterflies are unusual among insects in having two pairs of wings which function as a single unit, with the overlapping forewings providing all of the muscle power for the downstroke; this would not have been the case for the kalligrammatids and both fore and hind wings would have been powered during flight. Their flight was probably more similar to other large insects, without the erratic, floaty aspect of butterflies. However, at the end of the day, this is all pretty speculative, and it’s difficult to predict how they would have flown – there are no modern insects with comparably large wings except for the butterflies and moths, so there aren’t really any good modern comparisons. The two main takeaways from this would be that they lifter their legs up before take-off and their wings could be moved independently, and likely did not always function as a single unit.

 

Just a quick note on pollination and mimesis

I’ve already written far more than I intended to, and I’ve covered all of the main points I wanted to talk about, so here’s a few quick points to wrap things up.

Firstly, pollination and feeding. As I’ve mentioned at least twice already, most kalligrammatids had proboscises, similar to modern butterflies and moths. These were pretty diverse and varied in both length and how robust they were (Labandeira et al., 2016 supp. info.). In general they were long (8 – 20 mm) and differed significantly from the proboscises of other insects of the time (longer and more robust). Several families of Mesozoic plants are suggested to have been hosts for the kalligrammatids; the cycads, bennettitaleans, caytonialeans and potentially some cheirolepidiaceans (a family of extinct conifers). Among these the bennettitaleans, specifically members of the family Williamsoniaceae, have been suggested to be the most likely candidates for having an association with kalligrammatids. These were shrubby plants which produced a variety of flower-like reproductive organs which were often large and likely produced nectars (Fig. 5). Another interesting observation is that the proboscises of different kalligrammatids varied a lot, with some species possessing thin and gracile proboscises, while others bore longer, more robust proboscises; this suggest that some niche partitioning was present in the kalligrammatids, and different species likely fed upon and pollinated different plants (Labandeira et al., 2016).

Fig. 5 A kalligrammatid (Kalligramma brachyrhyncha) visiting a bennettitalean female cone (Wielandiella villosa) (Pott et al., 2015). Cutaway showing the internal cone structure. Note that the long robust proboscis is roughly the lame depth as the scales within the cone. Also, a cute toothy grin! Scale bar 10 mm.

Just a final quick note, before I end this far-too-long blog, is that at rest, the kalligrammatids would have folded their wings back over their bodies; this has several implications: firstly, when at rest the hindwings would have been completely or almost completely hidden by the forewings; secondly, the dorsal surfaces of the wings was the side visible when at rest (this is very different from butterflies, which fold their wings up together so that the underside of their winds is visible). This helps to explain why the eyespots of most species are present on the forewings, as they would have been hidden when at rest if on the hindwings. The undersides of the wings also may have been duller in colour, or at least may not have possessed such striking eyespots as the tops of the wings. It is also worth noting that the forewing shape in most species, and the large size of these wings roughly matches the size and shape of the heads of many small predatory dinosaurs, which are known to have inhabited the same biotas as the kalligrammatids. It seems likely then that the maniraptorans probably served as a model organism for the kalligrammatids. Furthermore, the large eyespots with round pupils and multiple rings closely matches what we predict the eyes of small maniraptorans were like (although, again, this is speculative; Fig. 6). While it seems likely that most kalligrammatids were mimicking their model organisms while at rest this does raise the question of why some species had eyespots of their hindwings. A possible, albeit speculative, explanation for this may be that species with hindwing eyespots would take flight when disturbed, suddenly exposing their previously hidden eyespots to would-be predators, startling them and scaring them off. I imagine this would be particularly affective in Makarkinia, with its massive wings and large eyespots.

Fig. 6 Kalligrammatid mimesis, showing the possible mimicry of small feathered predatory dinosaurs (maniraptorans).

That’s all folks

Well, that’s it for now, those are all of my thoughts on the kalligrammatids, based on the restorations I’ve recently been doing of them. Again, I just want to reiterate that these are mostly just my own personal thoughts and observations, but I hope that if nothing else this will help get people interested in the kalligrammatids and restoring some more of the amazing diversity which was present within this fascinating group.

Fig. 7 Oregramma illecebrosa about to take off from someones hand. Hopefully this will give you an idea of what a kalligrammatid would have looked like in life and this should also help to show just how large these insects were.

References

Ansorge, J., Markarkin, V. N. 2020. The oldest giant lacewings (Neuroptera: Kalligrammatidae) from the Lower Jurassic of Germany. Palaeoworld, vol. 30, pp. 296 – 310, doi: 10.1016/j.palwor.2020.07.001

Bechly, G., Makarkin, V. N. 2016. A new gigantic lacewing species (Insecta: Neuroptera) from the Lower Cretaceous of Brazil confirms the occurrence of Kalligrammatidae in the Americas. Cretaceous Research, vol. 58, pp. 135 – 140, doi: 10.1016/j.cretres.2015.10.014

Engel, M. S. 2005.  A remarkable kalligrammatid lacewing from the Upper Jurassic of Kazakhstan (Neuroptera: Kalligrammatidae). Transactions of the Kansas Academy of Science, vol. 108, iss. 1, pp. 59 – 62

Labandeira, Yang, Q., Santiago-Blay, J. A., Hotton, C. L., Monteiro, A., Wang, Y-J., Goreva, Y., Shih, C., Siljeström, S., Rose, T. R., Dilcher, D. L., Ren, D. 2016. The evolutionary convergence of mid-Mesozoic lacewings and Cenozoic butterflies. Proceedings of the Royal Society B, vol. 283, art. 20152893, doi: 10.1098/rspb.2015.2893

Liu, Q., Lu, X., Zhang, Q., Chen, J., Zheng, X., Zhang, W., Liu, X., Wang, B. 2018. High niche diversity in Mesozoic pollinating lacewings. Nature Communications, vol. 9, art. 3793, doi: 10.1038/s41467-018-06120-5

Lu, X-M., Zhang, W-W., Liu, X-Y. 2016. New long-proboscid lacewings of the mid-Cretaceous provide insights into ancient plant-pollinator interactions. Scientific Reports, vol. 6, art. 25382, doi: 10.1038/srep25382

Machado, R. J. P., Freitas, A. V. L., Ribeiro, G. C. 2021. A new giant species of the remarkable extinct family Kalligrammatidae (Insecta: Neuroptera) from the Lower Cretaceous Crato Formation of Brazil. Cretaceous Research, vol. 120, art. 104724, doi: 10.1016/j.cretres.2020.104724

Popa, M. E. 2019. Review of the bennettitalean genus Weltrichia. Journal of Palaeogeography, vol. 8, art. 12, doi: 10.1186/s42501-019-0023-9

Pott, C., Xiaoli, W., Xiaoting, Z. 2015. Wielandiella villosa comb. nov. from the Middle Jurassic of Daohugou, China: More Evidence for Divaricate Plant Architecture in Williamsoniceae. Botanica Pacifica, vol. 4, iss. 2, pp. 137–148, doi: 10.17581/bp.2015.04115

Yang, Q., Wang, Y., Labandeira, C. C., Shih, C., Ren, D. 2014. Mesozoic lacewings from China provide phylogenetic insight into evolution of the Kalligrammatidae (Neuroptera). BMC Evolutionary Biology, vol. 14, art. 126

Zhang, Y., Wang, X., Wang, S., Huang, W., Weng, Q. 2021. Kinematic and Aerodynamic Investigation of the Butterfly in Forward Free Flight for the Butterfly-Inspired Flapping Wing Air Vehicle. Applied Sciences, vol. 11, art. 2620, doi: 10.3390/app11062620

Zhong, Y., Huyskens, M. H., Yin, Q-Z., Wang, Y., Ma, Q., Xu, Y-G., 2021. High-precision geochronological constraints on the duration of ‘Dinosaur Pompeii’ and the Yixian Formation. National Science Review, vol. 8, art. nwab063, doi: 10.1093/nsr/nwab063