Violets and others: Rosa vosagiaca. Glaucous Dog-rose

Rosa vosagiaca. Glaucous Dog-rose, previously part of Northern Dog-rose pair.

Rosa caesia subs. caesia (Northern Dog-rose) and its other half Rosa caesia subsp. glauca (Glaucous Dog-rose) , have been split into separate species. Subspecies glauca is now called Rosa vosagiaca.

This split has made the identification of the most common hybrid rose in Cambridge R, caesia x R. canina = R. x dumalis more difficult, as there are now more combinations.

R. vosagiaca Great Massingham, Norfolk 29th Sept 2026

Most Dog-roses in Cambridgeshire are hybrids of either of these two Northern Dog-rose species.

The previous subspecies of R. canina (Dog-rose) was split into three separate species R. canina, R corymbifera and R. squarrosa and it has proved quite difficult to find 'pure' examples of each . My recent blog on R. canina goes through some of the problems, when trying to find a plant that can be photographed, to represent a good R. canina. By 'pure' I mean a plant that conforms to nearly all the required features as listed in the new BSBI handbook by Roger Maskew and Gareth Knass. Some introgression has to be allowed for as 100 percent purity would be very rare.

The two 'Northern Dog-roses' have introduced many features into local Dog-roses, however there are no examples of Northern Dog-roses themselves, south of a line from the Bristol Channel to the Wash.

I was keen to find and photograph a Northern Dog Rose so I could better understand its features.

This proved very difficult as 'UP North', in Yorkshire, County Durham and Northumberland I tracked down grid references for these species only to find plants that failed to have many of the required features. They were it seemed hybrids with mainly R. canina agg. My conclusion is that both species of Northern Dog-rose are a lot rarer than the BSBI maps indicate.

It was then great news that Lewis Saunders and Alan Leslie had found several bushes of R. vosagiaca in Norfolk. Samples were sent to the BSBI Rose referee, Gareth Knass who confirmed they were R. vosagiaca.

Follows are some photos of these plants found on a roadside bank at Great Massingham, West Norfolk.

R. vosagiaca and R. canina .agg 26 Sept 2026. on roadside bank.

First impression was a row of about 6 well established large rose bushes with high arching stems ran along the road bank for about 20m. They showed multiple very large orange hips, except the single R. canina agg which had smaller darker red hips.

First feature was to check, was the hip pedicel which was clear of glands in nearly all samples but very occasionally one or two stalked glands were present. This was considered acceptable introgression. The pedicels were very short.

R. vosagiaca. Typical group of hips with very short pedicels.

The key features of R. vosagiaca are the very short pedicels, large bracts, a large hairy head of stigmas which cover the disc, and a wide stylar orifice. The above photo shows the short pedicel and the bract, although somewhat damaged, is longer than the pedicel.

R. vosagiaca with a large head of stigma with hairs, covering the disc.

R. vosagiaca spit hip showing the orifice is about 1/3rd the diameter of the disc and the hairy stigma.

The above photo also shows that the disc is flat, which is a feature of R. vosagiaca.

R. vosagiaca, abundance of bright red hips.

Another important feature of R. vosagiaca is that the sepals are persistently erecto-patent at least until the hips are ripe. The above photo shows that the sepals are starting to fall off, because the hip is quite ripe although still firm. Just before falling off they are raised to an upright position.

R. vosagiaca Hips showing raised upright sepals just before falling.

R. vosagiaca, close up of sepals showing some stalked glands and some pinnate side protrusions.

Leaves.

R. vosagiaca. New growth leaf showing folded leaflets and slightly glaucous darker upper surface and paler underside.

R. vosagiaca. Leaf shape: leaflets well spaced apart, quite narrow and not very pointed at the apex.

Leaf shape in all dog-roses is known to be very variable according to Tony O'Mahony who has studied roses in Co Cork for many years. The above leaf shape certainly fits with the new handbook but details like margin type, glands and hairs are probably much more important than shape.

Leaf margin.

R. vosagiaca. Leaf margin: mainly uni-serrate with some partial bi-serrate tips.

No glands present except a few on the margin towards the base of some leaves and no leaf hairs, glabrous on both surfaces. All good as defined in the new handbook.

R. vosagiaca. Rachis has some stalked glandular hairs.

The handbook suggests most plants have no hairs and are eglandular or occasionally have a few scattered glands and pricklets. Maybe this indicates some introgression with a species that has glands.

Stipules. Not covered in the new handbook but Tony O'Mahony documented that Dog-roses have elongated glandular tips which fall off quickly, whereas Sweet Briars and Downy Roses have round glandular tips which stay on. See separate blog covering this feature, which works on over 95% of roses tested in Cambridgeshire.

R. vosagiaca.

The round glandular tips and fact they are staying on the margin, as shown above, suggests that some introgression with a non Dog-rose has occurred but since the only other variance with the expected set of features, is the sparse stalked glands on the rachis it can be accepted as R. vosagiaca. This was confirmed by Garath Knass who considered there might be a very small introgression with a Downy rose.

If I find this type of variance, I double check features, for example, do a more thorough check to see if any acicles are present, in order to see if an element of R. rubiginosa might be the culprit.

R. vosagiaca New shoot showing wine-red pigmentation.

R. vosagiaca Prickles on new shoot.

Typical dog-rose prickle.

Conclusion.

At long last I have taken a reasonable set of photos of one of my target species, R. vosagiaca. I thank Lewis Sanders and Alan Leslie for finding these plants and also arranging for the BSBI referee Gareth Knass to check out a sample. It was impressive that Gareth Knass came back quickly with an excellent critique and confirmed that he considered the sample to be R. vosagiaca.

A remaining question is whether these roses were planted or arrived naturally, as they are just outside the normal distribution range. There was no indication that they had been planted, except a bit further up the bank beyond a gate there was a single rose bush of a type of garden rose.

At some stage it would be nice to find a R. caesia but that might be next year now.