I so love the delicate flowers of Ayenia!! But it took time & effort for me to search out and study relevant literature, in order to more fully learn about & appreciate the details of its unusual floral morphology & biology.

Notably, I was confounded the very first time I saw the parachute-like flowers of Ayenia...because it appeared to me the stamen filaments were situated outside the corolla! [This initial (mis)conception may be fairly common...e.g. the second sentence in this post suggests a similar take.]

I was confounded because this seemed to violate a fundamental principle of floral organization. That is: the four constituent whorls of the standard floral form...namely, the calyx; corolla; androecium; and gynoecium...should always be concentrically nested in the ('outer-to-inner') order just stated. In some taxa, one or more of those constituent floral whorls may become reduced, vestigial, or even entirely lost through some particular evolutionary trend towards functional parsimony (e.g. to facilitate wind pollination; produce unisexual flowers; etc.). But…for whichever whorls that are present (and here, all 4 are)…the order is always compatible with calyx, corolla, androecium, gynoecium. [Presumably this is a consequence of persistence of an underlying floral framework inherited over 100+ million years of evolution from a primordial common angiosperm ancestor.]

So seeing what appeared to be ''stamen filaments outside the whorl of petals'' in Ayenia flowers had truly baffled me! But it turned out my initial perception was mistaken...the androecium whorl is indeed inside the corolla whorl. This became clear upon studying the descriptions & line drawings in various references [e.g. the Jepson eFlora and FNA treatments...and delving into further detail: Dorr(1996); Cristobál(2003); Cristobál(1993); Cristobál(1992); Cristobál(1969); and Machuca(2017)].

Here's how the tiny flowers of Ayenia compacta are organized (it may be helpful to open the full-size image in a separate browser tab and toggle back-and-forth while reading the following):

1. Calyx: An outer whorl of five, lance-ovate sepals emanate more-or-less perpendicularly from the outer edge of the tip of the pedicel. The sepals here are pale-yellow on the upper (adaxial) side and speckled-red below (i.e. abaxially)...with their edges mildly down-curved from the midrib


2. Corolla: Next, moving inward and alternating with the sepals, comes the whorl of five petals. Each petal starts at its base as an ascending, filiform, pale-whitish 'claw' (here slightly 'wriggling') and then turns reddish as it bends gracefully inward towards the central axis of the flower and abruptly flares into a widened, spade-shaped petal blade whose planar distal portion is held perpendicular to the central floral axis. The petals converge apically...with the two apical-edges of each spade-shaped blade lining-up almost seamlessly with the corresponding edges of the adjacent petal blades. Note also that the tip of each petal is subtly notched...and proximal to that notch, on the outward-facing (abaxial) side of the petal blade, is a curious little structure: a dark, gland-like, linear, and typically reflexed appendage. Overall, to the naked eye the distal edges of the petal blades appear fused into a shallow dome...but in actuality they are simply connivent (= ''just-touching'')...both with each other laterally, and apically with the top of the 'central column' of the flower...described next.


3. Androecium: Like the corolla, the next concentric whorl moving inward is quite distinctive in Ayenia. It consists of a relatively-long tubular stem emanating upward from the central base of the flower, called an ''androgynophore'', which...at its apex...flares into a bell-shaped structure with a reflexed rim, referred to as the ''staminal tube''. (Though, rather than a ''tube'', it's shaped more like an annulus centered around the style base...I think a better name would be ''staminal crown'', or ''androstelma'' for those who prefer Greek terminology!). Placed equidistantly along the upper/outer edge of the staminal tube are five very short, free-filaments...each with an anther dangling from its end (composed of 3 thecae). Alternating with those free-filaments & anthers along the rim of the staminal tube (= androstelma ;-) are five small broadly-triangular 'flaps', referred to as ''staminodes''. [For illustrations of all this, scroll to the bottom of Fig. 1 from Dorr(1996) to view 'item G' there; see Figs 1B, 1F & 1G from Cristobál(1992); and Figs 1B & 1D from Machuca(2017)...and see also Chris Christie's photo here, where the petals of the flower at center are splayed out to reveal a clear view of the staminal tube & androgynophore. Another interesting illustration is Fred Melgert's photo here (from midway down this image page), showing an apparent ''sport'' where the staminal tube has only 4 anthers.]
   
   Note that the apex of each petal blade overlaps onto the upper rim of the staminal tube, and the small flaps of each notched petal-tip clasp the rim of staminal tube, fitting snuggly around the base of the short free filaments below them, with the corresponding anther tucked under the petal tip (see Keir Morse's photo of a flower viewed from below). All this, again, reinforces the illusion of a fusion of the petal blades with...not only with each other laterally...but also apically with the upper edge of the staminal tube. The overall effect of this floral morphology...from the thread-like petal claws to the arching petal blades connivent with each other and the rim of the campanulate-bowl atop the androgynophore...gives the impression of an elegant miniature ''parachute''. [BTW, I can now better explain my initial mis-perception. In viewing the tiny flowers here, I had mistaken the central androgynophore & staminal tube as a corolla tube...and also mistaken the thread-like petal claws for stamen filaments (lying outside of that 'faux corolla tube'), and the linear appendages atop the petals as anthers! ;-]


4. Gynoecium: Finally, a very short 'free' style extends upward and out of the center of the bowl-like staminal tube atop the androgynophore. It supports at its tip a 5-lobed stigma, which ultimately receives pollen for fertilization.

• Note that in the typical configuration of Ayenia flowers, as described above, each petal blade is laterally connivent to its neighboring petal blades, and has its notched tip clasping the filament base if the anther opposite it along the rim of the staminal tube (photos: view from above; view from below). So the “canopy of the parachute” forms a physical barrier isolating the anthers (under the “parachute”) from the stigmatic surface (above the “parachute…in the center of the ‘bowl’ atop the androgynophore). Thus this configuration of Ayenia flowers constitutes a distinctive & strong form of herkogamy! Prior to the mid-1980's, herkogamy was mostly thought to be an adaptation serving to promote out-crossing, but Webb & Lloyd(1986) noted that many instances of herkogamy occurred among known self-incompatible taxa, and proposed that in such instances the more significant function of herkogamy was to reduce 'interference' to effective male & female function within a given flower (i.e. by reducing wasted self-pollen deposited on a flower's stigma, and wasted ovules due to 'stigma-clogging' and pre- or post-zygotic failure of self-fertilization of ovules, etc). This has now become a commonly-excepted notion (cf. the 'Introduction(s)' in Galloway(2010) and Li et. al.(2013)). And, indeed, Ayenia is currently placed within the tribe Byttnerieae of subfamily Byttnerioideae in family Malvaceae...which according to this 'APG Phylogeny Website' page is presumed to have a shared family character of post-zygotic self-incompatibility (a form of Late-acting Self Incompatibility).
• As seen, for example, in Chris Christie's photo here, and in my photos here and here…it seems that in some flowers (in later stages of anthesis?) the petal blades become detached and 'pull away' from the rim of the staminal tube. How is this “pulling away” accomplished? Might it involve movements of insect visitors triggering disengagement of the clasping 'teeth' at the apex of the petals, causing the petal blades to spring outward due to tension in the petal claw? Could this be related to an effective pollination behavior?
• I don’t know whether floral maturation in Ayenia is protandrous (= anthers shedding pollen before stigmas become receptive) or protogynous (= vice-versa…see the Wikipedia dichogamy page)…but it seems the interactions between dichogamy, herkogamy, petal blade movements, and the behavior of insect visitors may make for an interesting pollination story here!

In their entirety, the tiny flowers of Ayenia are an incredibly delicate, intricate, and delightful creation!

And we've only barely touched upon perhaps the most fascinating topic...i.e. how this floral structure presumably relates to the behavioral dynamics of Ayenia's insect pollinators, and the mutual influences flower and pollinator may have on each other...as well as their mutual evolutionary impacts over time?