Wednesday Weirdness: Cannonball Tree

I will admit that Pharyngula's Botanical Wednesday feature today induced a kind of a panic attack that only a botanist can have:

Couroupita guianensis, the cannonball tree

See that big, pink, sea-anemone-looking structure just kind of floating there? Well that's (1.) some sort of biological structure that's (2.) apparently part of a plant and (3.) I don't know what the hell it is!

To figure it out, I was forced to pay a visit to the font of all scientific knowledge: Wikipedia. Wiki was able to provide with an abaxial view:

Ok, so the sea-anemone is actually attached to the receptacle (i.e., the fleshy structure supporting the reproductive bits). Good. It would be pretty embarrassing if there was a plant with the power of levitation that I hadn't heard about.

The cannonball tree is bee pollenated but, unusually, produces no nectar. The bees come to the flowers looking for pollen (which is quite nutritious, as any health food store proprietor will tell you). In PZ's photo, you can see two sets of anthers: the set on the anemone-looking thing (hereinafter the 'hood') produce sterile pollen intended as a pollination reward. The less-conspicuous set below produce fertile pollen. As bees collect the sterile food pollen from the hood, they get fertile pollen all over their backs. This allows pollen to be transfered to the stigma (the thing in the centre that looks, I'm sorry, like a nipple) of another flower. It's a really cool pollination mechanism; I don't think I've ever seen anything like it in another flower.

The phylogenetic position of this thing is also quite interesting. Given that it has large flowers and lots of fleshy stamens, I would've expected it to be in the basal Angiosperms along with the water-lilies and magnolias. In fact, it's not even remotely related to this group: it's in the Ericales, an order that includes tea, blueberries, primroses and rhododendrons. This has evidently been an evolutionarily busy family.