British cuckoo bees – an aide-mémoire

Screen Shot 2017-04-07 at 09.47.39

The British bee season is well underway with lots of reports on social media of queen bumblebees (and even workers in the south), and male and female solitary bees (especially early emerging mining bees – Andrena).  In my own garden I’ve already spotted a couple of bumblebee species, plus the Hairy-footed Flower Bee (Anthophora plumipes) and the Grey-patched Mining-bee (Andrena nitida), amongst others.  Running alongside the emergence of these nest-building bees is a whole suite of “cuckoo” or “cleptoparasitic” bees that, as the name suggests, lay their eggs in the nests of other bees, consuming the pollen that has been collected and, usually, the eggs and larvae of the host bee.

The specificity of the interactions between the cuckoos and their hosts varies a lot.  Some are very host specific, such as the bumblebee sub-genus Psythirus that only parasitises other Bombus species.  Others are much broader in their host use, such as the genus Nomada that parasitises five other British bee genera.

Personally I struggle to recall which cuckoo bees interact with which host bees, especially for those with a broader use of hosts, so I thought I would construct an aide-mémoire in the form of an interaction graph using the R package “bipartite”.  I took the information on which cuckoo bees parasitise which hosts from Steven Falk’s recent (and very good) book Field Guide to the Bees of Great Britain and Ireland.  If anyone spots any errors, please let me know!

The bipartite graph is structured such that the hosts (to the left, in black) are ranked from most to least parasitised (in terms of number of cuckoo genera that interact with them).  The cuckoo bees (in grey on the right) go in the reverse order, from most specialised to least specialised.  Note that this set of interactions only applies to Great Britain and Ireland; breadth of host-parasite interactions is wider on the Continent and elsewhere in the world.

Here’s a link to a better quality PDF of the plot that you’re free to use for your own use: Cleptoplot

Here’s the data matrix (Clepto) and here’s the R script if you want to play with it:

> library(bipartite)

#Turns the CSV data file into a data frame and assigns the first column to be the row names

> Clepto2<-data.frame(Clepto, row.names=1)

#Basic plot of the web

> plotweb(Clepto2)

#To turn the plot 90 degrees and centre the image, change spacing and text size, colours, etc.

> plotweb(Clepto2, method=”normal”, text.rot = 90, labsize =1.5, ybig = 0.7, low.y = 0.7, high.y = 0.98, plot.axes = FALSE, y.width.low = 0.05, y.width.high = 0.05, col.high = “lightgrey”, bor.col.interaction=”black”, bor.col.high=”black”, low.spacing=0.03, high.spacing=0.08)

#Note: save the figure as a PDF, much better quality than PNG

#With thanks to Kat Harrold who provided some of the script

9 thoughts on “British cuckoo bees – an aide-mémoire

  1. ibartomeus

    You got me with bee parasites and networks in R! If I have time I would like to superimpose phylogenies to that graph. This may help remembering as Stelis, for example, parasite in Megachilidae, although I am surprised that most other parasites do not track phylogeny as nicely.

    1. jeffollerton Post author

      Yes, lots that could be done with this! Some really nice phylogenetic patterns. I started by thinking about doing it at a species level but there’s some degree of uncertainty as to which bee species are hosts/parasites.

  2. Ellen Dorothea Moss (@EllenDMoss)

    Very nice idea, I will certainly keep a copy with my keys and books. I had a quick go at ordering them by family, I find it easier to remember stuff like this if I can see what family they are in.

    ## Code:

    Net1=read.csv(“Ollerton_clepto_EDM1.csv”, header=TRUE, sep=”,”, row.names=1)
    Host=read.csv(“HostFamily_EDM1.csv”, header=TRUE, sep=”,”, row.names=1) #column of host genus names, column of host family names
    Para=read.csv(“ParasiteFamily_EDM1.csv”, header=TRUE, sep=”,”, row.names=1) #column of para genus names, column of para family names

    plotweb(Net1, method="normal",
    col.high=Pcolours, bor.col.high='black',
    col.low=Hcolours, bor.col.low='black',
    text.rot = 45, labsize =1.3, adj.high=c(0,0), adj.low=c(1,1),
    ybig = 0.7, low.y = 0.7, high.y = 0.98,
    plot.axes = FALSE, y.width.low = 0.05, y.width.high = 0.05,
    low.spacing=0.03, high.spacing=0.07)
    pt.cex=1.3, cex=0.8)

  3. Philipp Uhl

    Dear Jeff,

    very interesting post. In fact I liked it so much that I tried make a similar plot for German bee species. My source of information is Paul Westrich’s book “Die Wildbienen Baden-Württembergs” ( For every parasitic bee species where data was available the presumed primary host was extracted. The resulting dataset comprises of about 120 parasite species plus host species (out of appr. 500 German bee species). Afterwards, I created this plot using your part of your R code:

    I did not rank host species genera by number of interactions because the plot got to crowded. I had to use the “cca” method where a minimum of interaction crossings are created. However, one can still see which genera have more interactions from the size of the boxes. Since I had interaction data on the species levels this was also incorporated in the matrix and is therefore represented in the plot (I am not sure if this explanation is comprehensible).

    I really like that one can see some kind of clustering. Is this a result of co-evolution in certain genera?

    Many thanks to Nadja Schnetzer who extracted the parasite host interaction data from Westrich’s book.

    All the best


    1. jeffollerton Post author

      Dear Philipp,

      Thanks for this, it looks great! Interesting to see genera that we don’t have in Britain, such as Biastes, and the fact that they parasitise genera which over here are parasite-free, such as Dufourea.

      Yes, there’s clear clustering, and I think that some of it is phylogenetic and reflects the fact that cleptoparasitic genera are evolving from non-parasitic groups (Psythirus/Bombus being the obvious example). Did you look at Ellen’s plot? That shows it well by colour coding the boxes by family.




Leave a Reply to jeffollerton Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s