Tackling Diptera families – where to start and how to progress

In Britain, the fly fauna now comprises around 7,100 species. The number has risen dramatically in the last twenty years, by around 350 species. It is currently the largest single component of our fauna but may eventually be surpassed by the Hymenoptera if the taxonomy of the Parasitica is ever resolved!

The sheer numbers give us a lot of headaches. Can one individual really tackle them all? If you cast around it is clear that even the most competent Dipterists tend to specialise.

The big question is how to get started and how to progress from there onwards? The trouble starts with finding a reliable key to families; we don't have a published key that is simple. 

Life gets complicated because there are now two systems for naming the parts: for example the Comstock-Needham system of naming the wing veins, which has been around for a century or more, and McAlpine which is more recent. Older Royal Entomological Society keys use the former and later ones tend towards the latter. So, when you start with Diptera there is an added level of confusion with two completely different sets of anatomical names! We do make life difficult for ourselves!

Most Dipterists dislike the AIDGAP key by Unwin, and there is agreement amongst the leading specialists that it does not work in some places. The best readily available key to families is Pjotr Oosterbroek's key to European families of Diptera but it is not for the faint-hearted. There are many technical terms and unless you are familiar with, for example, the nomenclature of wing veins, it is often hard work switching back and forth from key to illustrations.

Stuart Ball and John Ismay have been working on a key that deals with the British fauna. It is now pretty well developed, but cannot be published because it relies very heavily of illustrations cribbed from other publications. Stuart is in the process of photographing all of the relevant features so something may emerge eventually.

So, where to start?

Most people tend to start by noticing animals that are obvious. Leaf-baskers and flower visitors. They naturally gravitate towards families such as the Syrphidae for this reason. But it is not just Syrphids that visit flowers: lots of Calyperates do too, especially Tachinidae, Calliphoridae and Muscidae. Unfortunately, user-friendly modern keys are not available for all of these families and modern entomologists are somewhat spoiled by the keys to Syrphids. Stubbs & Falk does an amazing job of turning a family that was once considered too hard for all but the museum specialist into one that can be tackled with relative ease. [I do stress relative ease – parts of this family are far from straightforward!]

The big advantage of Syrphids is that once you have learned what they look like, you have also been introduced to quite a few non-Syrphids and a certain amount of comparative anatomy – not a hoverfly but what is it? If it has bristles then a lot of people are rapidly turned off because the keys rely on the relative positions of the bristles or bristle scars. Nevertheless, once you have mastered hoverflies you may want bigger challenges. Soldierflies and their allies sit comfortably with hoverflies so they too get tackled quite early on. Thereafter, it is a question of whether there are accessible keys but, perhaps more importantly, also whether you can actually use the key and have some confidence that the point you arrive at is reliable.

In my early days running ecological surveys I had some interesting discussions with colleagues who took the view that the end point they had reached was always right – whereas I felt that one should check further and remain cautious. If there is a species description, read it! If the description says that your species is confined to the far north of Scotland and you have recorded it in Dorset you are as likely as not wrong! We see an awful lot of duff data that can be eliminated quite quickly because the geography is wrong.

Moving on

In my early days as a Dipterist, there was a lot of interest in hoverflies because Stubbs & Falk had just been published and even the seasoned 'experts' were breaking new ground. Over time, those people have moved on from hoverflies and into other families. The challenge is breaking new ground and finding species you've not seen before – it is a sort of 'collector' approach that is utterly understandable: unless you are interested in some form of data interpretation the greatest thrill is something new.

Today, many of those former hoverfly enthusiasts tackle other families and only take a passing interest in hovers. But, they do this as a progression: there is a moderately workable key to Dolichopodids and they are quite attractive flies; Male Empis and Rhamphomyia, with their exhuberant genitalia, are also interesting and Collin's key is very workable (in parts); there is a reliable key to Sciomyzidae and they are eminently doable; the Tephritidae are pretty doable (but I do have trouble in places), as are Conopids, Otitidae and Uliidae.

Thereafter one starts to enter a minefield. There are keys to Phorids, but I would not touch them (very few Dipterists will) – they require slide mounting, as do Psychodidae. Likewise very few people tackle Sciarids or Sphaeroceridae. More importantly, when starting with a new family you really don't have the markers that help you find your way around the key. If you have access to a reliably determined collection then you can get to grips with a new family; if not, how do you know whether you have got the right ID? Your fly is pink with blue spots, but all the key does is to tell you that it has two notopleural bristles and crossed post-ocellars! You are none-the-wiser!

Where the keys are sparsely illustrated and lack species descriptions  potentially doable families are as yet under-attempted. The Tachinidae are a case in question. They are often big and seemingly obvious, but try using the key! Unfortunately there are a lot of single species genera that make the key a long procession that becomes very confusing. I've yet to master it but have this on my to-do list for the winter! I've been collecting Tachinids all year and now have three or four hundred to tackle. Hopefully that will get me further!

Tackling a new family

It is not really viable to tackle a key with just a single specimen or indeed a handful of specimens. Within moderately large families you just don't have the comparative material needed to understand what the key is talking about and even some small families can be a problem without relevant comparative material.

The system a lot of the more adventurous Dipterists use is to collect for a couple of seasons before attempting to get to grips with the key. When I do this, I attempt specimen one and see how I get on? If I hit a block then I put it to one side and try another, and another and another. Each time I familiarise myself with another facet of the key and start to see how it is constructed and how the writer has interpreted the morphology. Over time come a few small successes: those start to be the markers in the key – I know what that is and the next specimen is not it! We try to do the same when running training courses – make sure that critical parts of the key are embedded and the student has some simple markers to work from.

Crucially, this all takes time. Finding your way around the families, recognising features such as costal wing breaks and head chaetotaxy requires infinite patience and access to comparative material. So, the follow-up is a need to maintain a collection. That in turn becomes the limiting factor – store boxes and cabinets are expensive, take up a lot of space and require curation, so in the end you have to specialise!

Making envelopes for papering gnats and craneflies

If you are minded to collect either (or both) craneflies or fungus gnats for Alan Stubbs and Peter Chandler, it is simple enough to store them in advance of sending them off. Making the storage system is described in six simple steps.

1. Cut a square of paper about 7cm square (or thereabouts). I usually fold over an a4 sheet into two and then two again - this gives me potential for 12 squares. Alternatively, you could use old postage envelopes and simply cut off the corners to appropriate sizes.

Figure 1. Basic square of paper

  2. Fold the square into two to make a triangle

Figure 2. Folded into two to make a triangle - in this case the square was slightly rectangular but that does not matter.

3. Turn over one open side of the triangle to seal it. I usually don't worry about using tape to make sure it stays closed but you can if you feel so inclined.

Figure 3. One side of the triangle sealed by folding over.

4. Write relevant data on the triangle BEFORE filling it with specimens otherwise you will crush them.

Figure 4. Triangle with data - I usually smaple both gnats and craneflies so I need to put a note as to what the envelope contains.

5. Fill the envelope with sorted specimens - you should be able to open the envelope by putting pressure on two sides to cause the remaining open side to part open. However, to show what the envelope contains I have opened one to show how the flies are arranged.

Figure 5. Triangular envelope containing fungus gnats.

6. Seal the open side by turning over and making sure there is a good fold.

Figure 6. Sealed envelop ready for storage.

It is probably best to liaise with Peter and/or Alan before sending large volumes of specimens, but both of them are usually pleased to get specimens that will improve coverage by their schemes.

Scottish square-bashing: an update

When I went to Scotland in June, I made a serious effort to record craneflies and fungus gnats in addition to hoverflies. Wherever there was suitable habitat for these families, and it coincided with a potential recording site for hovers, I swept around for a little while. On most occasions my efforts were quite limited, but occasionally I spent a reasonable time sweeping because the weather was not conducive to recording hoverflies.

I've just had a summary of the fungus gnat results back from Peter Chandler. It transpires that I managed 221 species/site records from 30 sites comprising a total of 94 species (about 17% of the British fauna). Unfortunately, my timing was not ideal and I was probably a little early in the year but I did get a few whose distribution is mainly Scottish. The best site list I generated was at Craigellachie where I managed to find 42 species which is almost a respectable total!

Possibly the best record I generated was Ditomyia fasciata from a site near Barnard Castle in Co. Durham (see map of records up to 2011). Apparently I also took a specimen of this species in Wingate Plantation during my visit to John Bridges in June. These two records are clear northward extensions of its known range. Maybe it is responding to climate change but, then again, coverage is not fantastic in NE England (although we did have a summer field meeting based in Durham in 2005).

Figure 1. Distribution of the fungus gnat Ditomyia fasciata to 2011

Peter tells me that fungus gnat recording has been very limited this year – our spring field meeting generated a few records and the Summer field meeting generated about as many records as I generated in Scotland, but with about 30% fewer species (one set of specimens still to be processed may change things).

This feedback is very useful because it highlights just how important parataxonomists such as me could be to this recording scheme. Peter lives in southern England and will never manage to cover the country on his own. So a combination of field meetings and individual efforts is needed to improve coverage. So, do you own a net and a microscope (and pooter)? If so, it is not too late to start collecting fungus gnats and supplying them to Peter.

I’m now hard at work sampling Mitcham Common for its gnat fauna. It is darned hard work as there do not seem to be many gnats about. Perhaps it is the season; but, then again, the site supports relatively young woodland and is infested with bramble so it may just be that it is not a good gnat site. Time will tell.

How mobile are hoverflies and other Diptera?

When out walking I make a serious effort to record the insects that I come across. Obviously most, but not all, are hoverflies. On several occasions I have seen relatively uncommon saproxylic species well away from woodland. Two species immediately come to mind: Myolepta dubia, which I found in several places last year, and Mallota cimbiciformis which I found once in a tree-lined avenue. I recall, however, recording Mallota at Thrislington Plantation (an open grassland) a decade or more ago. At the time I thought it odd because there were no decent-sized trees for at least a mile in any direction.

Casting the mind wider, last year I found the snipe fly Chrysopilus erythrophthalmus ovipositing on an isolated Black Poplar where it had been felled and in the same area found the cranefly Gnophomyia viridipennis. Both species breed in decaying sap, a highly transient habitat. They must have arrived within a year of the tree being felled: so where did they come from?

Do we need to develop a register of mobility attributes?

Once one starts to think about insect mobility, the question then arises: do we actually know much about the mobility of insects of conservation interest? I think we are just starting to scrape the surface. Work on the Aspen Hoverfly Hammerschmidtia ferruginea has shown that it is actually comparatively mobile. Again, I have encountered this species substantial distances away from potential breeding sites. By inference, I think we can also assume that Callicera rufa is very mobile, given that it is now cropping up at a wide variety of English localities.

Conversely, there are other species that appear to be anything but mobile, or they are so specialised that those individuals that stray never find suitable habitat? Examples that come immediately to mind include Caliprobola speciosa that is seemingly confined to the New Forest and Windsor Great Park, and the Pine Hoverfly Blera fallax that is restricted to two closely approximated localities in Speyside.

Where are, and which are the intermediates?

Relevance to conservation management

When I worked in the Invertebrate Site Register Team back in the 1980s, the thinking was that saproxylic species were relatively immobile. After all, if they flew too far, they would end up in a wilderness of open countryside! That approach was clearly wrong! Or at least was it? It seems to me that we need to think in much more detail about insect mobility from a conservation management perspective. Many invertebrates utilise micro-features that are transient to some extent and so they must move as the situation demands. The crucial point is that the landscape needs to support sufficient of these micro-features as a constant asset, even if the individual features appear and disappear.

There are of course the flagships for meta-populations such as Marsh Fritillary Butterfly. But, who is thinking about strategic distribution of micro-habitats for other invertebrates? Conservation strategies tend to focus on big projects – major site developments. Perhaps one of the most useful programmes might be to work on relatively small features that are important and distinct micro-habitats? For many years I have felt that there was a case for locilised reduction in grazing pressure over base-rich seepage lines. Perhaps too, there is a case for developing a strategic vision of a landscape that will eventually be richer in veteran trees with associated micro-habitats?

I also wonder if there is a need to develop projects that investigate the availability of important invertebrate micro-habitats. Not large-scale mapping of macro-features, but a log of important micro-features such as trees with different types of rot and their distribution across the landscape. Can we make any correlation between densities of such attributes and the distribution of important saproxylic assemblages? If there is a link (and I'm sure there is one) then the best way to improve the overall resource of saproxylic species is strategic planting to secure veteran trees across hitherto uninhabited landscapes.

Changes in the numbers of contributors to the Hoverfly Recording Scheme

In the second provisional atlas published in 2011 we commented that just 20 people had contributed over 50% of the data to the HRS. At that point the number of photographic contributors was small but growing. Data from photographs largely came from Flickr and iSpot, and relatively few people contributed more than a few records each. In the intervening six years the landscape of biological recording has changed markedly, with a far bigger constituency of contributors.Stuart recently produced a nice illustration that shows some of the changes (Figure 1). For this analysis recorders have been defined as people who have submitted at least 50 records AND have made submissions in at least two year.

Figure 1. Numbers of recorders who have contributed 50 or more records over two or more years. The lower plot shows the number of such people per year from 1979 to 2014. By definition, "recorders" have to have been around for at least two years before they could be counted so the count is likely to be lower in the most recent years - hence the drop at the end. The top plot shows the number for which this is their first year of submission - i.e. new recorders.

These plots got me thinking about how the numbers work in terms of yearly contributions and the numbers involved in the HRS, year on year. Today, the total number of contributors exceeds 8,000 names going back to the 19th Century. I suspect that there will be some duplication where people use multiple aliases that I have not detected, but I doubt that this is significant. In recent years, the number of contributors has exceeded 1,000 individuals per season, but until the advent of the Facebook group many of these would have been single records from a photo posted on Flickr or other hosting systems. In the last couple of years the number of contributors has dropped quite markedly because I no longer find it time-efficient to search Flickr during the winter. So, the numbers of single entry contributors have dropped.

In order to get a feel for what is happening I assembled the data in three ways:

• the number of recorders in each year who, between them, contributed 50% of the data for that year;
• the total number of contributors in each year; and
• the average number of records per contributor.

These graphs tell an interesting story:

Figure 2. The number of recorders whose combined efforts generated 50% of the records in a given year between 1975 and 2016.

Figure 3. Total number of contributors in each year between 1975 and 2016.

Figure 4. The average number of records per contributor between 1975 and 2016.

What these three supplementary graphs tell us is that there have been several events in the life of the HRS and that the composition of the recording community has changed.

In the 1970s and early 1980s records came from a very narrow spectrum of Dipterists, most of whom would have used microscopes and specimens and would either have relied upon Coe's key or possibly test keys for the first edition of Stubbs & Falk.

The publication of Stubbs & Falk led to an increase in recording activity and the arrival of a small number of very active recorders whose contributions meant that the average number of records per recorder jumped quite markedly. Several of that cohort continue to be active to the present day, with recruitment of new recorders relatively constant at around 12-15 new people per year until around 2006. This is the point in the process where Stuart and I started to ramp up our training programme and I think there is some evidence that it has had an effect. This also about the point where digital photography started to gain popularity and a new cohort of photographic recorders entered the system.

By 2008, the impact of digital photography is pretty clear. We cannot claim the rapid growth in contributors to have resulted from our courses. At best we have trained about 500 people in the past ten years and I would be amazed if more than 20% of trainees ever go on to do any serious recording (probably many fewer). So, the real changes have come from the advent of iSpot, Flickr and, latterly, the UK Hoverflies Facebook group.

The dip in the average number of records per recorder around 2011/2012 coincides with the period of greatest activity by me scanning Flickr for photographs that could be converted into records. Lots of single record recorders were created as a result. Meanwhile, the climb in the average over the following four years suggests a substantial change in the composition of the community with many more people making a serious effort to record hoverflies and to post their findings on the Facebook group. We can see this in the steep rise in the number of contributors making up 50% of the incoming data.

What is perhaps less obvious is that there are vastly more people who contribute 100 to 500 records per year. Figure 5 reveals just how this cohort has changed, and it is clear that the UK Hoverflies Facebook page has had the most dramatic impact on recording.

Figure 5. The numbers of contributors of 100 or more records in each season from 1975 to 2016.

What is especially noteworthy about Figure 5 is the degree to which numbers are comparatively constant between around 1987 and 2013. The big change has taken place from 2014 onwards and I think reflects a growing skill-base and interest that can only have come from the Facebook group. To my mind this demonstrates the potential of such media and the importance of engaging with wildlife enthusiasts through modern social media rather than by traditional routes.

Autumn field meeting - Tarbet - some reflections

For more than 40 years there have been Autumnal field meetings organised, primarily, to survey the cranefly and fungus gnat faunas of the British Isles. Initially they were run by Alan Stubbs in his role as Head of Entomology at the Nature Conservancy Council. Upon Alan's retirement in 1991 there was no capacity to organise these meetings and they became a voluntary effort. Peter Chandler ran them for many years but decided to give up the role in 2003. I therefore stepped in and have run them in the subsequent years. My reason for doing so was not a great love of craneflies and fungus gnats, but instead a feeling that we would be losing a valuable event that has played a big part in our understanding of the autumnal fauna of the British Isles.

Initially, I followed the preceding model, which was a four-day event in mid to late October. The weather at this time is rather variable, so the dates have moved towards mid October starting on the second Saturday of the month or around 14 October (whichever is closer to the middle of the month). I also changed the format so that we spend a week in the field. My thinking is that there is that a lot of time is spent travelling and it is generally not practical to visit more northerly locations for just three or four day's fieldwork that might get rained off for part of the time. So, we now tend to go for a week, and often use two centres in order to improve coverage. Figures 1 & 2 show where we have been since 2004.

Figure 1. Autumn Field meeting coverage 2004 to 2012

Figure 2. Autumn field meeting coverage 2013 to 2017

Participants in the Autumn meeting are a fairly tight-knit group - very few people want to visit wet dank woods at the end of the season! So, in most years we only get six to eight participants. Also, we tend to attract people with interests in a wider range of disciplines; not just Diptera. I think this is very positive because the team makes a much broader contribution to the data on autumnal insects in far-flung parts of the country.

In a further change, I introduced the idea of supplementary meetings in late August or early September to cover Scotland where Autumn comes a lot earlier. Thus, in 2017 we found ourselves gathered at the Bay Tarbet Hotel on the west side of Loch Lomond. There were six of us: Peter Chandler (Fungus Gnat Recording Scheme), Alan Stubbs (Cranefly Recording Scheme), Andrew Halstead (Symphyta and many other Orders), Keith Alexander and Janet Lister (Saproxylic Coleoptera and Bark Bugs) plus me. On these trips I act as a parataxonomist and simply hoover vast numbers of Nematocera that I then sort into separate piles for Alan and Peter. I usually retain a residue of a few families such as Heliomyzids and Lauxaniidae and will of course log hoverflies when I come across them.

On this trip, hoverflies were comparatively abundant (figure 3) so I did assemble a reasonable number of records. Nevertheless, it is clear from the data that a small number of species made up the bulk of the records and that there was a comparative dearth of interesting species. That does not matter because coverage of common species can be useful and we did get a very good picture of what is flying at this time of year.

Figure 3. Numbers of records of hoverfly species recorded during the DF field meeting based at Tarbet from 8 to 15 September 2017.

In the course of six days (we lost one day to rain) we managed to visit 21 10km squares within a 50 mile radius of Loch Lomond (Figure 4)  and got a good feel for the fauna of the area. We visited a great many more 1km squares, often having to stop at two or more points within individual 10km squares in order to get anything like an adequate sample. Thus, the data I already have assembled covers more than 40 data points. Stopping points can rarely be considered to be sites; rather, they are points where it was possible to park the car and access suitable habitat. We tend to look for wooded streams and wetlands, as these are the most suitable for gnats and craneflies.

I think some of the ecological reflections are best left to a further post, as there is a need to try to interpret some of the results. Not all of the material has been worked-up and Peter has a large volume preserved in alcohol. Alan has done most of his specimens and tells me the species list is a bit over 50 species of cranefly and associates. There is one possible highlight - I took a specimen of a Tipula that Alan is unsure about. It may prove to be something that is already known but there is an outside chance that it is an addition to the British list - I wait in eager anticipation. Thankfully it is a male so it ought to be identifiable.

Figure 4. Locations visited during the 2017 DF meeting based at Tarbet. Note the three more southerly squares are visits made on the journey north on 7 September.

Understanding Eristalis: some simple rules

Eristalis is one of the most frequently encountered hoverfly genera, with two species (E. tenax & E. pertinax) occurring almost entirely throughout the year. Four further species are common from April until October (E. arbustorum, E. hortiicola, E. intricaria & E. nemorum. The remainder are scarcer (E. abusiva & E. rupium) or vanishingly rare (E. cryptarum & E. similis).

Although widespread and often abundant, they are far from easy to identify but a few simple rules help to arrive at a diagnosis on many occasions. Nevertheless, even the most experienced specialist will stop short of making a firm diagnosis on those occasions where critical characters are unclear. It is important, therefore, to be aware of the limitations of what can be done and to err on the side of caution if critical characters are not adequately depicted. So, where should one start?

The first rule is to avoid making a diagnosis based solely on abdominal patterns or wing shades. Eristalis are infinitely variable in both respects and so these characters are rarely reliable diagnostic characters on their own. The complications arise because there is seasonal and sexual polymorphism in many species. Attention needs to be paid to the colour and shape of the legs, the extent of dusting on the face and the shape of the stigma on the wing. The other useful feature is the size of the animal, but judging this from photographs is always problematic so it is not terribly helpful in many cases where a photograph or series of photographs is involved.

The one exception is Eristalis intricaria which is a bumblebee mimic and is most frequently confused with Volucella bombylans and perhaps Merodon equestris. Unlike the latter two, E. intricaria has partially yellow hind tibiae and the scutellum is paler (the colour under the hairs and not the hairs themselves).

Leaving aside Eristalis intricaria, we need to think about a logical sequence of diagnosis. So, where to start?

There are two species with nice obvious characters that are usually visible if a sequence of photographs is taken from various angles: the colour of the front feet and the shape and colour of the hind tibia. So, to get started, we can say:

1a. Hind tibia completely dark (may be slightly paler towards the junction with the femur). Hind tibiae slightly curved and with feathering. Eyes with a strong band of dark hairs (visible in most good photos). ……………………………….Eristalis tenax

1b. Hind tibia partially yellow towards the junction with the femur (usually about 25% of the tibia). Eyes lack dark hair band …………….couplet 2

2a. Front and mid tarsi yellow or orange (beware occasionally muddy but never fully darkened on any segments). ……………………………….Eristalis pertinax

2b. Front and mid tarsi with some segments dark…….. couplet 3

3a. Wing with stigma tightly quadrate – not diffusing or extending towards the junction of vein R1 with the Costa. Face with a distinct un-dusted central stripe …………Eristalis nemorum

3b. Wing with more extended stigma …………………….Couplet 4

4a. Face completely dusted, hind metatarsus enlarged and stigma diffusing towards the junction of vein R1 with the costa …………………………….Eristalis arbustorum

4b. Face with obscure undusted stripe or with a distinct un-dusted stripe ………. Couplet 5

5a. Wing with a strong area of infuscation (clouding) that is often blackish in places …… Couplet 6

5b. Wing may be somewhat clouded, especially around the veins but lacking distinct clouds. This leads into the problem area where we often cannot be sure what we have.

6a. Hind metatarsus dark. Wing cloud distinct but not extensive …….. Eristalis horticola

6b. Hind metatarsus pale yellowish. Wing cloud often extensive and dark …….. Eristalis rupium

From this point on, life gets very complicated.

Some Eristalis arbustorum have partially rubbed faces and making sure this is rubbing and not a facial stripe is often a problem. Separating E. arbustorum from E. abusiva is very tricky and is dependent upon the length of the hairs on the arista and the colour of the middle tibia (partially darkened in E. arbustorum and almost clear yellow in E. abusiva). We do see exceptional photographs that capture these characters nicely, but they are the exception and so many of these examples only get as far as Eristalis sp.

Very occasionally, we see specimens of what appear to be quite large animals with dark feet but no band of eye hairs and a very distinct pale base to the tibiae. If the thoracic pleurae are well depicted, and the hind femur is clearly shown, then sometimes we can detect ashy dusting of the pleurae and of the hind femur – Eristalis similis; but beware – confusion with (especially) E. nemorum is possible and I have seen quite eminent Diptersts make this mistake (and have done so myself).

Finally, there is Eristalis cryptarum. In theory this should be easy to identify from photographs because it has completely orange hind tibiae. In practice, I'm not sure that it would be recognised as an Eristalis from some angles and on the one occasion I encountered it in the wild I did not recognise it immediately! This one needs careful checking because it is confined to a very small part of southern Dartmoor!

In making a diagnosis, it is always worth bearing in mind that there are further species in Europe that complicate matters. Although we do not expect them to occur in the UK, they might just be here and undetected; that would complicate matters very seriously!