This summary is based on the work of Perret et al. in their original taxonomic work and its subsequent update. I have tried to extract the relevant data without all the details; all mistakes in the extraction are my fault.
It is noteworthy that two types of molecular genetic analysis did not agree completely, and that the synthesis of the two did not place some species where they might naturally be expected to wind up. This disagreement with our expectations is hardly surprising. For one thing, our expectations might well be incorrect. For another, the molecular genetic analysis might lead to the wrong conclusions in some cases. Ideally, the raw DNA sequences are converted to a tree by generating all possible trees and then choosing among them by a process called maximum parsimony, a statistical version of Occam's Razor. However, for a genus as large as sinningia, the number of possible trees is reeeeeally big: it is larger than the number of atoms in the known universe. Therefore a simplifying technique called bootstrap analysis is used to generate and evaluate some likely trees. (Actually, there are a number of decision-making techniques, which don't always give results in agreement with one another.)
This procedure is very reliable but not infallible. Thus, if the results do not agree with our expectations, we can hope for further work to determine whether the problem is our expectations or the limitations of molecular phylogenetics.
(I also drew up a comparison of the three data sets used by Perret et al, and how they did as far as matching our intuition about relationships in the Sinningiae.)
That being said, and with the possibility that some details of this arrangement of species will change in the years ahead, this is how the species of Sinningia and its two related genera Vanhouttea and Paliavana seem to be organized.
Among other things, this list can give some hints for possible hybridization. Here's a hairy one: Sinningia hirsuta x Vanhouttea lanata!
Note: in what follows, clade means a group of plants which constitute all the living descendants of a common (and usually assumed to be extinct) ancestor.
The five main groups (with the Clade names used in Perret et al.) are
The molecular data suggest that the Corytholoma and Sinningia clades have a common ancestor not shared by the other clades, and then these two have a common ancestor with the Dircaea clade not shared with the fourth and fifth clades. That is, the evolution went something like this:
__________ Dircaea clade | | | ____ Corytholoma clade T___ | | | |_____| | |____ Sinningia clade _______ | | | | |_______________ "Free calyx lobes vanhouttea" clade Original ____| sinn | | |________________________ Thamnoligeria clade
The first two clades contain only tuberous species. The fourth and fifth clades have no tuberous species. The sinningia clade contains both tuberous and nontuberous species.
The authors state that the "acquisition of tubers in Sinningia has occurred independently several times". It's not obvious to me why they drew this conclusion. My amateur guess (based on nothing more than Occam's Razor) is that the tuber was invented by the common ancestor of the first three clades (marked T in the diagram), and that the paliavanas and vanhoutteas in the Sinningia clade are secondarily tuberless.
Those who can't resist this sort of thing can see some more diagrams and some logic and a bunch of naive arithmetic on the subject.
_____ Douglasii Group | ____| | | | |_____ Galea Group | ________| | | | |__________ Perennial stems Original ____| Dircaea | | | __________ S. eumorpha and S. conspicua | | |________| | |__________ S. lineata and S. macrostachya
Most of the species in this clade have been crossed with one another to form fertile hybrids.
Six members of this group -- marked with a (D) -- have douglasii-type streaked flowers. Flowering type is noted for each species.
S. nivalis and S. polyantha are closely allied species, although widely separated in habitat (mountainsides vs. seashore). S. piresiana and S. rupicola are also closely related (but the former is much easier to grow than the latter).
The three species at the bottom of this table (S. canescens, S. macropoda, and S. bullata) are fringe members of the group, but don't form a group of their own.
|S. leopoldii||terminal cluster|
|S. polyantha||(D)||extended axis|
|S. nivalis||(D)||terminal peduncle|
|S. piresiana||(D)||terminal cluster|
|S. rupicola||(D)||terminal peduncle|
|S. insularis||terminal peduncle|
|S. striata||(D)||extended axis|
|S. douglasii||(D)||terminal peduncle|
|S. calcaria||terminal cluster|
|S. leucotricha||terminal cluster|
|S. canescens||terminal cluster|
|S. macropoda||terminal peduncle|
|S. bullata||axillary cyme|
This group of species all have a galea. The upper two lobes of the corolla are combined and extend well beyond the bottom three, to form a kind of hood, beneath which first the anthers and then the stigma project out of the flower tube.
It is a little surprising that this group consists precisely of the sinningias with a pronounced galea. Taxonomical groups defined by a single character have often not held up to scrutiny.
Other genera (e.g. Columnea) also exhibit galeas. A few other sinningia species (e.g. S. araneosa and S. valsuganensis) have flowers with projecting upper lobes which are shorter and not fused into a galea.
See the galea page for more discussion.
In the 2001 classification, S. reitzii is sister to the rest of the Dircaea clade, which implied that all the other species in the clade are more closely related to each other than they are to S. reitzii. If S. "Black Hill" is a separate species, it might well be S. reitzii's closest relative.
The 2013 presentation by Alain Chautems at the Toronto convention of the Gesneriad Society revised this placement. According to the more recent data, the species S. reitzii and S. mauroana, with perennial stems, are more closely related to the above groups than all of them are to the two following groups.
This group and the next one are "sister" to the rest of the Dircaea clade.
S. eumorpha and S. conspicua certainly look closely related.
_____ Core group | ____| | | ___| |_____ S. harleyi | | ____| |__________ S. nordestina | | | | ______| |______________ S. aghensis + miniatures | | | | | |___________________ S. barbata Original ___| Corytholoma | |__________________________ S. richii
Most sinningia species grow within a hundred miles or so of the Atlantic coast of Brazil. However, some species of this clade, S. allagophylla, S. elatior, S. incarnata, S. sceptrum, and S. warmingii, have a wider distribution. S. incarnata is found as far north as southern Mexico. These species have expanded out of the typical sinningia habitat of rocky areas with moderate to high rainfall, into drier, plains areas.
This group includes most of the common and widespread members of this clade. There are two main subdividisions, but they must be closely related. S. incarnata, from the second subgroup, forms fertile hybrids with several members of the first subgroup, and S. sceptrum (2nd subgroup) yielded a fertile hybrid when crossed with S. aggregata (1st subgroup).
It is reasonable to hypothesize that all members of the Core Group form fertile hybrids with one another.
Sinningia 'Apricot Bouquet' is a fertile hybrid involving three species in this group.
This group is the geographical champion of the Sinningieae. S. incarnata has the widest distribution of any sinningia species, followed by S. elatior and S. sceptrum. S. incarnata and S. elatior are the only species found north of the Equator. All three are relatively tall.
It's worth noting that the nuclear DNA puts the tall species S. elatior, S. sceptrum, and S. incarnata in a group with other tall species like S. curtiflora and S. warmingii, rather than with S. defoliata and S. sp. "Canastra", but that the distance between them is still quite small.
Here we have five isolated groups, of one or two species each. One might guess that S. bragae will turn out to be a member of the S. aghensis group. In the first version of this page, I wrote that, "It would not be surprising if a future refinement of the data placed S. concinna close to S. pusilla." The latest update does that.
S. richii is the the outlier of the Corytholoma clade, not closely related to any of the other species or groups.
Sinningia barbata is sister to the rest of the Corytholoma clade minus Sinningia richii, which implies that all the other species in the clade (apart from Sinningia richii) are more closely related to one another than they are to S. barbata (see the clade tree).
Sinningia richii is sister to the rest of the Corytholoma clade, which implies that all the other species in the clade are more closely related to one another than they are to Sinningia richii.
___N________ Vanhoutteas | | __T____ S. gigantifolia and S. cochlearis |____| | |__N____ V. fruticulosa | |___N________ Sinningia gesneriifolia ____| | |___T________ S. hirsuta | | | |___T________ S. lindleyi | ________| | | | | __T____ Speciosa/guttata group | | | | | ____| | | | | | | | |__N____ Two paliavanas | |____| | | | |___T________ S. tuberosa Clade _______| Father | | |__N_______________________ Two more paliavanas T = branch with all plants tuberous N = branch with all plants tuberless
This clade is a jumble of tuberous and tuberless species. Recent molecular work has only increased the jumbling. This suggests that, in this group at least, acquiring a tuber or losing one or both are relatively easy to do.
Not very many hybrids have been made among the species in this clade. S. 'Amizade' is S. kautskyi x hirsuta. Plenty of varieties and "hybrids" of S. speciosa have been created. These exploit variety within S. speciosa itself, but may also involve some hybridization with related species such as S. guttata.
The 1988 AGGS Sinningia Register lists the hybrid 'Betty Stoehr' as being from a selfing of S. eumorpha x S. speciosa, which would imply that the primary cross was fertile. I'm skeptical.
One thing to note here is the distance between S. guttata and S. lindleyi. These species bear a distinct resemblance, particularly in the leaflike calyxes which surround the flowerbud. The nuclear DNA suggested that they were closely related, but the chloroplast DNA (inherited only in the seed-parent line) put them quite far apart, as indicated below. It will be interesting to see how this plays out in the future.
The type species of Sinningia, Sinningia helleri, is most likely in this clade, given its similarity to Sinningia guttata. Unpublished work in Geneva apparently indicates that this is true.
S. gigantifolia and S. cochlearis are closely related species. They have two notable features in common. First, they have alternate leaves, instead of the opposite leaves found in all other sinningias. [Although I have to say that my plants of S. gigantifolia have mostly opposite leaves.] Second, when winter comes, their stems die back to a point one or two nodes above the previous year's die-back, so that the stems gradually extend.
Vanhouttea fruticulosa is a tuberless subshrub.
This recently rediscovered species has "free calyx lobes" and would seem to belong in clade 4, but the molecular data say otherwise.
The subgroup it belongs to has four tuberous species and six nontuberous species.
Sinningia lindleyi resembles S. guttata in several ways, particularly the leafy calyxes, but while the data puts them both in the Sinningia clade, it does not put them close together within that clade.
These vanhoutteas (and one paliavana) have sepals (calyx lobes) which do not adhere to one another in bud (unlike Vanhouttea lanata and Paliavana tenuiflora, for example, where the calyx completely encloses the bud -- see a comparison).
In his presentation at the 2013 Gesneriad Society convention in Toronto, Alain called this the "Vanhouttea" clade, dropping the "free calyx lobes" modifier, perhaps because of Sinningia gesneriifolia, which has free calyx lobes but belongs to the Sinningia clade.
This group comprises only three species, including both sinningia species which don't form tubers. None of them flowers abundantly under my conditions, although S. schiffneri is the best of them. It shares with paliavanas the habit of developing long bare stems, sending up side shoots from the base, but S. gerdtiana is freely branching when young.