A start into Ferns. Bipinnate or Tripinnate.

A start into Ferns. Bipinnate and tripinnate ?

I started looking at ferns in the autumn of 2020 in West Cork as my normal summer trip had been delayed due to COVID and flowering plants were over their best.

As a complete beginner I made some progress helped by the arrival of the new fern book, 'Britain's Ferns' by James Merryweather which arrived towards the end of my holiday. This new book contains a lot of information and many photographs and is well worth the price of £16.  It does lack some of the detail found in 'The Ferns of Britain and Ireland' by Chris Page which was published in 1982 and has more on hybrids etc. No photos though.

This blog looks at Buckler ferns, as I made basic mistakes in 2020 and missed identifying Broad Buckler- fern ( Dryopteris dilitata).  This was corrected in 2021.

The first reason for going wrong was not understanding the distinction between bi-pinnate and tri-pinnate, which is used a a basic key differentiator in 'Britain's Ferns'. Second reason was not fully appreciating the difference in the size of the pinnule teeth of the Shield ferns vs. the Buckler ferns. 

As a complete beginner, the sheer density of information in 'Britain's Ferns' key was a problem.  Not fully understanding the definition of bipinnate and tripinnate definitely puts you at a disadvantage, along with the fact that this is not a clear distinction in practice.

The above diagram is based on the textbook definition of bi and tri pinnate (Ref.  The Kew Plant Glossary) and is not to scale.  Nor does the shape in the textbook diagram come close to a real fern but it does show the principle.

 If each division were stalked as shown, by the blue Pinnule and purple Pinnulets, it would be fairly straight forward but in real life ferns have a much more complex shape.  In a real fern, the divisions might start off being stalked but become merged, so the distinction between bi and tri has to be based on looking at the lowest pinna rather than the tips. This applies to the pinna division as well as the further pinnule and pinnulet divisions.  An example of these problems is shown in the green highlighted pinnules which show that the pinnules are clearly stalked at the base of the pinna but become merged at the tip. Also the pinnule are segmented into pinnulets at the base of the pinnule but become merged towards the tip.

 The amount of segmentation that forms the pinnulets can cause problems which will be illustrated later using photos.

One suggestion might be that it pays to be methodical. Start with the main frond stem (stipe) and moving up from the base, determine what is a pinna, the first division.  Then move up that pinna and check if the pinnule nearest the stipe is divided into pinnulets, i.e.  separated into distinct sections. This is where it can difficult, as this third division is not always clear cut and this gets worse as you travel up and out on the frond. Also as I will demonstrate later, not all plants conform to the segmentation being complete, even at the base.

Buckler Ferns.

Buckler ferns are tripinnate in 'Britain's Ferns' initial Key. The pinnulets have pointed tips but not as strong as those shown on the bipinnate Shield Ferns. Broad Buckler-fern has the classic shuttlecock form. Separation of the different species of Buckler-fern is difficult but the Broad Buckler-fern normally  has stipe scales that have a dark line through the middle. Separating Buckler ferns species is not the goal of this blog, but it was these dark strips that first alerted me to the correct identification. I had been seeing a Buckler-fern not a Shield-fern. As a beginner with no experience, you can make this type of mistake.

Broad Buckler-fern stipe near the base showing dark centre line.  23Sept21

  Back to the frond detail.

Detailed photo of the pinnule of a Broad Buckler-fern. 23Sept21

I might be breaking my rule to look at the lowest pinna but in this example even the mid pinna show a pinnule (shown outlined in red) has at the base pinnulets, which are segmented right down to the stem of the pinnule (outlined in white). Although the pinnulet base is quite wide and gets wider as you go up the pinnule the segmentation remains 100% until you get close to the tip. This is clearly a tripinnate frond. 

Broad Buckler-fern. Part of pinna near top of frond.

The above photo from the same frond but further up, shows that the pinnule is segmented but only to about 60%.  If you only looked at this part of the frond, the plant might be regarded as bipinnate.  

My first error using the Key was thinking Broad Buckler-fern was bi-pinnate because I was looking too high up the frond.

I wondered if I had found a Shield-fern, never having seen a Shield-fern?

Photo showing pinnulet tips follows.

Broad Buckler-fern showing tips to pinnulets.

Where it starts to get difficult, is that on some plants only the very lowest pinna and the closest pinnule to the stipe has anything like a full segmentation, all the other pinnules have partial segmentation.

There seems to be a lot of variation in this feature. Some fronds are clearly tripinnate for at least the lowest four pinnae and only above that become bipinnate.

Narrow Buckler-fern ( Dryopteris carthusiana) showing lowest pinnule lacking segmentation into pinnulets. 

The photo above shows the lowest pinna base coming off the stipe. The pinnule is segmented but only the very lowest pinnule next to the stipe is fully segmented right down to near the stem of the pinnule.

Photo of the lowest pinna base with the pinnule closest to the stipe. It shows the first pinnulet

is segmented back almost to the stem, the second segmented to about 90% and the third only to about 80%.  I have highlighted the pinnulet margin in black to show this. The next potential pinnulet out is far less segmented. I do not care if the pinnulet is attached on a broad base only its division from the next.

 I think the implication of this is that Broad Buckler-fern and probably all the Buckler ferns are variably bipinnate to tripinnate. 

Buckler-fern frond with tripinnate pinnulets highlighted in yellow, all the rest is bipinnate.

Conclusion. It is good to understand the principles behind the terms bipinnate and tripinnate but also be aware that in real life, ferns require a more complex description. It is interesting that Poland and Clements in their 'Vegetative Key to the British Flora' do not use the term bipinnate and tripinnate but use terms like 'Group BH - Lvs 2-3(4)-pinnate for all the Buckler-ferns, Shield ferns and even Bracken.  

I take from this that the Buckler-ferns can be bipinnate through to tripinnate and is therefore not a good fundamental separator to use in a key, as used in 'Britain's Ferns'. This applies if you are a complete beginner.  The experts know all this anyway.

To finish-  a few shots of Soft Shield Fern to show the different frond structure and the stiff terminal spines to the tips of the pinnules.

Soft Shield-fern, taken West Cork, 25Sept21

Soft Shield-fern. Base of pinna with pinnules with a thumb. 

Soft Shield-fern. Stalked pinnule with thumb and stiff terminal spines.

The first comment on Soft Shield-fern is that the pinnule is only slightly segmented so this is a bipinnate  frond. The shape is very different to that of a Buckler-fern with a much less symmetrical pinnule, unbalanced by the thumb which has a margin that runs very close to the stem.  The tips have terminal spines which are longer than those on the Buckler-ferns, but this is a variable feature.   I saw my first Shield-fern after returning from Cork in 2020, in late November in a ditch in Cambridge.

References:-

Beentje, H. 2020 The Kew Plant Glossary. (2nd Edn.) Kew Publishing.

Merryweather, J. 2020   Britain's Ferns. Wild Guides

Page, C.N. 1982 The Ferns of Britain and Ireland. Cambridge University Press.

Poland. J & Clements. E.J. 2020 The Vegetative Key to the British Flora.(2nd Edn.) John Poland.

 

Peter Leonard

Rampton

November 2021

Other terms not used in this blog:-

 The segmentation of the pinnules is partial and this is called pinnatifid where the depth is not specified, pinnatipartite where segmented to about half way and pinnatisect where divided to almost the midrib.

Narrow Buckler-fern  D. carthusiana.

 Yellow areas are tripinnate pinnulets which are separated to more than 90% of pinnulet length. Rest of frond is bipinnate.

Broad Buckler-fern D. dilatata 

Yellow areas are tripinnate pinnulets which are separated to more than 90% of pinnulet length. Rest of frond is bipinnate. There is a lot of variation about the extent of the frond being tripinnate.

These two fronds were collected in the Kings Forrest in February 22.

Rosa Mollis Soft Downy rose

Rosa Mollis  Soft Downy Rose,  a search up North....

Having started last year to photograph wild Roses found in Cambridgeshire, a chance to see some more northern species in 2021 seemed a good idea. The Downy Roses are a difficult group and with photos taken of R. tomentosa and R. sherardii in Cambridgeshire, an obvious target was Rosa mollis, the Soft Downy Rose.  And I wanted to photograph the most 'pure' example , if any rose is ever 100% 'pure'.

Since my daughter is now living in Newcastle, research of the BSBI maps showed potential targets near Newcastle and also around Allendale Town, home of the other mollis, Crepis mollis the Northern Hawks-beard. I then moved on to Teesdale and Ingleborough.

First stop was the Gosforth Nature Reserve just to the North of Newcastle, an area of wetland and woodland. Having paid the £5 entry fee,  I headed down the track from the entrance hut and within 20meters found the target, next to a Dog Rose. Five pounds well spent!

R. mollis, 3rd July 2021, Boundary Ride, Gosforth Nature Reserve.

 The first thing to notice was the dull leaves, a feature shared with the other two Downy roses. More subtle, was the lack of arching branches, giving it a lower height despite some thick near vertical stems. Flowering was nearly over but a couple of flowers that were still showing, were a bright and quite deep pink. Time to get out the BSBI handbook on roses and check that its features actually all conformed.

According to the handbook, the first key feature of R. mollis is the straight patent prickles. Prickles can be variable on roses and don't always quite obey the rules, so it is worth checking more than one stem to get an overall impression.

Mature prickle on an older stem, is patent (sticking out at 90 degrees) and quite thin.
It has a large oval base and is not unlike the lowest example shown in the illustration in the handbook.

R. mollis. Patent, straight and thin prickle. 3rd July 2021

R. mollis, New growth prickle.

R. mollis, stem just below pedicle with narrow fine prickles.

These prickles all look good for Rosa mollis as they are a critical feature and later, on my journey, I found R. mollis which were not quite so pure and started to have more curved and thicker prickles.

The main feature is that they are very thin despite having a wide base and that they stick out straight at either 90 degrees or slightly upwards. A comparison shot of Sherard's Downy Rose follows which shows more tapering and curving. 

R. sherardii. New growth prickle. 9th July 2021.

Lealets.

R. mollis. Upper leaf which is quite dull and leaflets just about spaced apart.

R. mollis. Upper surface is hairy. Quite dense simple white hairs all over surface.

R. mollis. Underside of leaflet. Tomentose 

This photo does not quite show the tiny glandular hairs which are almost hidden by the thick white simple hairs. These glands are hard to see due to their small size.

Blown up section of previous photo which just about shows the glandular hairs on the veins and leaflet surface. These are much smaller than the glandular hairs on the leaflet margin.

 These small sized glandular hairs are a key separator from the Sweet Briar roses like R. rubiginosa which have glands twice the size at 0.1 to 0.12mm dia.

Next up is the leaflet margin which should be 'irregularly glandular-serrate' according to the handbook.

R. mollis. leaflet margin.

The margin is certainly irregular with a basic uni-serrate shape but with secondary tips and glandular hairs either forming tips of their own or just stuck on the edge. The handbook description seems perfect.

R. sherardii  Leaflet margin for comparison is rather similar.

R. mollis Petiole and rachis have both short simple white hairs and glandular hairs.

R . mollis. Leaflet stipule has glandular hairs along the margin. They are very dense and different sizes with the larger ones, round tipped on short stalks that are more or less patent. The glandular tips are round not elongated.

Flowers

R. mollis. Deep pink colour and stigma cluster covers disc.

R. mollis. Stigma cluster covers disc and is about as high as it is wide, forming a dome.

R. mollis.  Sepals are almost simple with few tiny lobes and dense glandular hairs on the outside surface. Small leaf like tips to some sepals. Hypanthium and pedicle with stalked glandular hairs.

R. mollis. Sepals can be raised to an erect angle.

Pedicles are reported to be 0.5-1.5 cm in the handbook although the photo above shows one at about 2cm. Pedicle length always seems to be quite a variable feature and is supposed to be shorter in mollis than sherardii which has pedicles 1-1.5cm.     

With the above photos to confirm the details I was happy that this example at the Gosforth Nature Reserve is a good example of a Rosa mollis.  However I was lucky to start with such a conforming example and as I travelled on to Allendale and Teesdale the problems of hybrids became much more difficult plus the possibility of R. sherardii and its hybrids.

One plant at Teesdale looked good and  the a few more photos follow below:

R. mollis at Forest-in-Teesdale near the river.

Two key points here are, that the stigma cluster covers the disc and the sepals are almost simple with very limited lobes. This plus the low growth pattern without the arching stems are needed to confirm a R. mollis.  Quite a few plants were found that did not meet this test but still showed many features of mollis.

R. mollis at Forrest-in-Teesdale 

 Same plant as above photo showing quite a short pedicle with long stalked glandular hairs that are also on the hypanthium.

R. mollis type prickles present.

Also nearby were plants that had white flowers but otherwise did seem to tick all the mollis boxes.

R. mollis with white flowers. Forrest-in-Teesdale

Close up of flower showing the stigma cluster covering the disc.

Sepals almost simple, normally entire without side lobes .

The handbook does say that white flowered forms of R. mollis can occur and at Teesdale it would appear that both the deep pink and white forms occur.

Problem plants. The following are a few photos of plants that did not have all the required features.

Allendale Town plant.

The stigma cluster does not cover the disc. This and the fact that the prickles were not quite thin enough, suggested this was not a 'pure' mollis.

Allendale Town plant.

The plant was also too tall.  I am guessing maybe this plant is a hybrid possibly with R. canina but the real difficulty is that the shift in features gets close to sherardii which also has a slightly smaller stigma cluster and more curved prickles.  Sherardii is also a taller plant.  With all roses the problem of hybrids makes identification difficult. 

Allendale Town plant.

One feature that suggests a hybrid of R. mollis and R. canina rather than an example of R. sherardii is that the sepals are still quite simple without many lobes.  This plant would need expert attention well beyond my skill level.

Another problem plant was found in Allendale which had a lot of the features for R. mollis but did not have the right growth pattern with arching stems. It also had white flowers. This plant is very tall and arching stems are not correct for R. mollis.

Not the required shape.

Many mollis type features present but sepals have more lobes than expected. Also sepals
are not as erect as they should be and the hypanthium is not as round as might be expected.
 All pointers to not being 100% mollis. 

Conclusion.

To record a Rosa mollis is not easy. It has features that are close to R. sherardii, so care is needed to make sure all the main features are correct to exclude hybrids and sherardii.  I was lucky that my first plant which I found at Gosforth,  had all the features, as described in the Rose handbook.  Later at Allendale I found more plants, but some were certainly not 'pure'. The only other species of rose present in the area that I found, was R. canina agg.  so I suspect the mollis that had a more canina growth pattern were hybrids with canina. 

My search for R. caesia Northern Dog Rose was not successful and it is interesting that there are few recent records for this species and one recent BSBI record was, when located,  not accurate as it did not have many of the required features.

Hopefully the above photos are a good representation of what a Rosa mollis should look like and enable others to identify this species.

The main feature of R. mollis are:-

1) Growth pattern, low (2m max) with older plants having near vertical thick stems but without arching stems. Often suckering forming dense thickets (sherardii can also produce suckers).

Allendale. R. mollis, low thicket. 

2) Dull leaves, a feature shared with tomentosa and sherardii. Hairy on both surfaces.

3) Prickles that are thin and do not become much wider as you go to the base. The base is large, as if it supports a more normal canina type prickle. Prickles stick out at 90 degrees and are straight or almost straight. Not dense. If you find well curved prickles that is not good for R. mollis.

4) Sepals are almost simple without lobes ( actually single simple lobes are allowed) plus according to the handbook, can have leaf like extensions. Though I did not find that feature on the plants I found. These leaf like extensions are quite common in many species of rose, where the sepals turn into leaf like extensions suggesting that sepals and leaves are closely connected. 

5) Stigma cluster covers the whole disc or at least most of it whereas sherardii only covers 2/3rds. Both dome shaped unlike tomentosa which has a higher width to height ratio. Stylar orifice of R. mollis and R. sherardii is large at 1.5-2.7mm dia. whereas R. tomentosa is small at 0.4-0.8mm dia. and can be checked by removing the styles. Mollis should have a larger orifice than sherardii and size and shape of the orifice is an important feature and would have been worth investigating.  

Pedicles are short but there is much overlap in this feature with sherardii. One thing I have learnt over the last year is that to determine a rose, you have to check all the features. Some plants tick nearly all the boxes but one feature might not be correct and then you have to re-check with more concentration on the details and take a bigger sample from different stems. Make sure it's the same plant, not two different species with interwoven shoots, which can also confuse.  

Peter Leonard

Rampton Cambridgeshire

11th July 2021

Updated 21 July 2023 with following photos.

Rosa mollis orifice more than half the width of disc. 21July 2023

R. sherardii Also has a wide orifice but more like a third than a half. Also the sepal has a pinnate

lobe on the right and two tiny lobes on the left. This would be unusual for mollis which tends to have zero lobes but the handbook suggests some have leaf-like tips which would complicate this feature. 

R. mollis but shows variation in that its orifice is closer to two fifths rather than a third. This plant was possibly not quite showing the correct shape leaving just a little doubt, although all other features looked good for mollis. In practice there would seem to be some variation in the size of the orifice and possibly some overlap between mollis and sherardii. The good news is that R. tomentosa has a narrow orifice.

Bract extending into leaf.

R. mollis. Stipule gland fringed with round globose heads on short stalks.
Typically one large than a smaller one, then another larger one in quite a dense array along the margin of the stipule. No sign of the Dog Rose type of forward pointing elongated heads. 

Update July 2023 with extra photos.  

R. mollis sepals showing leaf like extensions which is noted as a feature in the Handbook no7.

16th July 2025 Grosmont, North Yorkshire. 

R. mollis hips with short pedicels and erect more or less simple sepals with leaf like extensions.

16th July 2025. The sepals were noticeably thin and upright at this stage of fruiting. Note that R. sherardii sepals can also occasionally have leaf like extensions but in my limited experience the sepals in sherardii sepals are wider and can have side lobes.