Epilobium Hybrid Key

Source: Kitchener, Geoffrey. “Epilobium Hybrid Key.” Sussex Botanical Recording Society Newsletter, no. 65 (January 2008). http://sussexflora.org.uk/wp-content/uploads/2016/03/Newsletter_Jan_2008.pdf.


Whilst rummaging through my bits and pieces I came across a rather tatty piece of paper that turned out to be a key to the Epilobium hybrids. Knowing that the expert on this genus is Geoffrey Kitchener I thought that it could be his work so I contacted him to seek permission to reproduce it in our newsletter. Not only did he agree to us using it, but he kindly updated it and adapted it for use in Sussex. It is here presented below with grateful thanks to Geoffrey Kitchener.

Arthur Hoare


This is for the most part deliberately not cast as a true key.  It would not be impossible to construct one.  Indeed, all known Sussex hybrids, bar one, are covered within a key by Smejkal (1997).  But any such key will always face the challenge of suggesting certainty where the extent of variation is too great for this.  The following steps are recommended in analysing potential hybrids.

Does the plant look like a hybrid, e.g. floriferous, perhaps willowy branches, semi-abortive capsules, seed at least in part shrivelled and sterile, flowers perhaps flushed purple, and the plant as a whole not clearly corresponding with a species?  If so, for semi prostrate plants outside Sussex, see section 2 below.  For erect plants, see section 1.  Willowherbs are notoriously variable, so do not draw conclusions too rapidly from a plant just because it does not correspond to the typical form of a species.

1.     Erect plants

Is the stigma regularly 4-lobed (see section 1.1); entire (see section 1.2); or misshapen like a clenched fist, or with irregular lobing (see section 1.3)?

1.1      4-lobed stigma. Parents are amongst these: hirsutum, E. lanceolatum, E. montanum, E. parviflorum. Features which these spp. may contribute are:

~E. hirsutum: large deep purple flowers, long stem hairs, half-clasping leaves, hooked teeth to leaves.

~E. lanceolatum: cuneate leaf base, longish leaf stalks.

~E. montanum: rounded leaf base.

~E. parviflorum: felted leaf surface, sessile leaves, medium length or longish spreading stem hairs.

All these spp. may contribute glandular hairs, but if in quantity, they are likely to come from E. hirsutum and/or E. parviflorum.  The most likely crosses in Sussex are E. parviflorum x montanum and E. hirsutum x parviflorum, followed by (as yet unrecorded) E. hirsutum x montanum.  Crosses with E. lanceolatum are unlikely because of the relative infrequency of that species in Sussex.

1.2      Entire stigma. Parents are amongst these: ciliatum, E. obscurum, E. palustre, E. roseum, E. tetragonum. Features which these spp. may contribute are:

~E. ciliatum: glandular hairs in quantity, ridged seeds.

~E. obscurum: smooth, appressed hairs, glandular around collar below sepals.

~E. palustre: narrow leaves, long seeds (in hybrids, where fertile, these will be longer than 1.3 mm), some glandular hairs).

~E. roseum: cuneate leaf base, long leaf stalks, glandular hairs in quantity.

~E. tetragonum: narrow leaves, smooth, finely appressed hairs, none glandular.

The most likely crosses in Sussex are E. obscurum x ciliatum or E. tetragonum x ciliatum.  These can be difficult to tell apart if all three species are present.  Other possible Sussex crosses, as yet unrecorded, are E. tetragonum x obscurum, E. obscurum x palustre and E. palustre x ciliatum.  Crosses with E. roseum are unlikely because of the relative infrequency of that species in Sussex, although E. parviflorum x roseum has been seen; also E. obscurum x roseum has been recorded twice, but one of these records has recently been re-determined as E. roseum x ciliatum.

1.3      Misshapen (confused) stigma. This includes plants that have mixed stigmas, some entire or 4-lobed and some misshapen.  One parent will come from the 4-lobed stigma group (section 1.1 above) and the other from the entire stigma group (section 1.2).  See above for contributory characteristics.  The most likely cross by far is E. montanum x ciliatum.  Otherwise, E. montanum x obscurum, E. parviflorum x obscurum and E. parviflorum x ciliatum may relatively frequently be encountered.  E. hirsutum x ciliatum, E. hirsutum x palustre, E. hirsutum x tetragonum and E. parviflorum x tetragonum are also known from Sussex.


2.      Semi-prostrate plants.

This section is included for completeness, but its application to Sussex is limited

2.1      Upland or montane (E. brunnescens absent): Parents may include E. alsinifolium, and/or E. anagallidifolium.  Not applicable to Sussex.

2.2      Lowland or upland (E. pedunculare present), plant with ovate well toothed leaves bronzed on underside: E. montanum x pedunculare.  Not applicable to Sussex.

2.3      Lowland or upland (E. brunnescens present), reddish plants, beginning as prostrate: E. brunnescens hybrids. These are very unlikely in Sussex, not just because E. brunnescens is very infrequent, but also in that hybrids appear to be restricted to Cornwall, Wales and Ireland.  They may be keyed out as follows (subject to the caveats above as regards the limitations of keys for hybrids).


1     Stems with hairs projecting more than 0.2 mm from surface. E. parviflorum x brunnescens

1     Stems with hairs projecting less than 0.2 mm from surface. 2


2     Stems semi-glabrous below, with lines of hairs descending from nodes; stigmas entire.3

2     Stems hairy all round below, stigmas “confused”.5


3     Leaves generally narrow; seeds (when fertile) 1.0-1.3 mm long with distinct neck where hairs are attached. E. palustre x brunnescens

3     Leaves ovate-lanceolate to elliptic; seeds (when fertile) 0.6-0.95 mm long, sometimes with necks where hairs are attached. 4


4     Glandular hairs present, often abundantly, along ovary at anthesis, extending at least to pedicel; fertile and sterile seeds with ridging on surface; seeds (when fertile) may have a slight neck where hairs are attached. E. ciliatum x brunnescens

4     Occasional glandular hairs present along ovary at anthesis, not extending to pedicel; seeds without surface ridging or neck; fertile seed surface tuberculate, sterile seed surface reticulate. E. obscurum x brunnescens


5     Flowers whitish in bud, white or pale pink in flower. E. lanceolatum x brunnescens

5     Flowers purplish-pink in bud and in flower. E. montanum x brunnescens




Smejkal, M. (1997). Epilobium L., in Slavík, B., ed. Květena České Republiky 5: 99-132. Academia, Prague