Centaurea nigra vs Centaurea debeauxii

Recorders' conundrums

Many of us know about C. debeauxii (Chalk Knapweed), but few of us have the confidence to identify it. There are very good reasons for this. These same reasons apply equally to the identification of C. nigra, and I hope this short summary of the issues will change the way we approach recording of both taxa in Sussex.

Stace only separated it from C. nigra (Common Knapweed) in the third edition of his Flora (2010). In the third edition of Flora of the British Isles (1987) C. nigra was divided into two subspecies: C. nigra ssp. nigra and C. nigra ssp. nemoralis, the latter becoming C. debeauxii, named after Jean Odon Debeaux (1826-1910). 

Our problems arise from the fact that the process of speciation is recent (and some would say incomplete), that the two taxa interbreed freely (both have 2n=44 chromosomes, despite Sell & Murrell stating C. nigra has 2n=22), and that a third species, C. jacea (Brown Knapweed), also participates in hybridisation with both species and their hybrids. 

This leaves us with the following possible taxonomic allocations: 

C. debeauxii                                                                  C. nigra

C. debeauxii x C. jacea                                                C. nigra x C. jacea

C. debeauxii x C. nigra                                                 C. nigra x C. debeauxii

                                  C. nigra x C. debeauxii x C. jacea 

Much of our understanding stems from crossing experiments conducted in the 1950s by Marsden-Jones & Turrill (1954), however an excellent paper in the New Journal of Botany by Margaret Harris (169-181. Vol 7; 2-3: 2017) has disentangled the practical aspects of field identification. The paper is free to access for BSBI members here, but I will try to summarise the main points.

The issue of separation has been discussed elsewhere. For one point of view which contains nice photos of the phyllaries see Ken Adams’ impressive website.

The most reliable characteristic is the degree to which the phyllaries overlap. The phyllaries have a scarious apical portion separated from the basal portion by a slight waist. These are edged with fimbriae. The degree of overlap can be categorised as:

1: None    2: touching fimbriae     3: partial overlap of fimbriae        4: regular overlap fimbriae onto neighbouring phyllary 5: complete coverage of underlying bracts

One must look at the third row of phyllaries from the base of the terminal capitulum of at least ten plants in a cluster and allocate a score to each. Add them, divide by ten and multiply by 100 to get the overlap index. Below 200 and you can be confident of C. debeauxii. 200-300: C. debeauxii x C. nigra. 300 – 400: C. nigra x C. debeauxii. Above 400: C. nigra. 

What about C. jacea? Well, despite no records of the taxon this century, its genes persist and can be recognised by the degree of fusion of the fimbriae:

1: none    2: few fused     3: most fused at base       4: most fused, some to halfway    5:  some fused more than halfway    6: most fused more than half way    7:  margin almost entire

Again, the same method should be used as for phyllary overlap. Any score greater than 150 suggests a C. jacea hybrid. 

The following chart summarises these points:

Calendar

Description automatically generated

Fig. 1. Boundaries of Centaurea taxa and their hybrids as defined by fimbriae fusion and overlap hybridity indices. (From A morphological analysis of Centaurea nigra and its associated taxa in Hertfordshire M. Harris. New Journal of Botany 2017: 7 p 169-181. With kind permission.)

Other characters given in field guides can be useful pointers, especially capitulum width, sub-capitulum swelling of the stem, and shape of the central undivided part of the apical part of the phyllary. By themselves they do not allow taxonomic determination as they are affected by environmental factors. 

This series from Mick Crawley takes a simpler approach.

It is likely that most of our plants (as elsewhere) are hybrids (or, as those who don’t accept the separation of species would say, intermediates). On the chalk it is likely that most are C. debeauxii or C. debeauxii x C. nigra (C. debeauxii s.l.). Contrary to the account in our Flora, C. nigra s.s. may actually turn out to be the least common taxon in Sussex. 

Some plants have capitulae with pseudo-radiate heads. This is more often seen with C. debeauxii. Stace doesn’t apply subspecies status to these as they usually occur amongst non-radiate plants, but they have been attributed thus by others: C. debeauxii ssp. thuillieri and C. nigra ssp. rivularis respectively. So how to record these plants? Our preferred name is Centaurea nigra sens. lat. (=nigra/debeauxii) when there is uncertainty. This is available on iRecord. If you feel able to allocate the plant more accurately please use C. nigra s.s. or C. debeauxii s.s. but mention the reasons in the comments section.