How small could a flower get? In a previous post I mentioned New Zealand's smallest flower, Wolffia arrhiza. The smallest New Zealand flower that has a corolla and so looks like a flower however, is considerably bigger (Myosotis brevis, at 1 mm diameter). But what's the smallest a flower could be?
Most flowers are hermaphrodite, but there are plenty of flowers that are either male or female. But for the sake of argument, let's say our flower must be hermaphrodite. It needs a pistil and at least one stamen, but maybe it doesn't need petals (corolla) and sepals (calyx).
The pistil is the female part of the flower. Let's keep it as simple as possible, a single carpel with a single ovule (the ovule becomes the seed). The carpel is a modified leaf, and its wall needs an epidermis on each side (say 10 µm thick), and a minimum of 1–2 layers of mesophyll in the middle, cells that are probably at least 15 µm ; that's a layer 4 cells in thickness (say 50 µm ) wrapped around the ovule. The ovule can have a single integument if it's an Asterid (other flowering plants have 2 integuments), maybe 2 cells in thickness. Inside that is a single layer of nucellus (megasporangium wall), and inside that an embryo sac of 7 cells. The whole ovule could be shrunk down to about 100 µm (the lily ovule featured in first year textbook Raven Evert & Eichhorn is about 400 µm across). The ovule needs a vascular supply, but it it's orthotropous (erect on its stalk) the thickness of a vascular bundle doesn't need to be added to the flower's diameter. So the single ovule surrounded by carpel wall could be about 200 µm diameter.
The stamen could similarly be reduced. If this tiny flower is cleistogamic, maybe it doesn't need complex walls like a lily stamen has (200 µm thick). Maybe 50 µm is enough, like the carpel wall. The smallest pollen grains are about 10 µm , and even a cleistogamic flower will need more than one of them. Firstly, because pollen grains develop from spores, and spores are produced by meiosis, 4 would be a theoretical minimum (without abortion, such as happens in the ovule). Secondly, even a habitual self-pollinator needs an excess of pollen grains to ovules, maybe 20–30:1 at a minimum. The internal volume of the anther couldn't be much smaller than about 30 µm diameter; adding the wall on both sides would make it about 130 µm across.
Such a flower doesn't have to have the pistil and stamen packed side by side. If the ovary of the pistil overlapped the filament of the stamen, and the anther of the stamen overlapped the style of the pistil, they could fit together into a smaller space. The resulting flower is about 230 µm diameter, a pretty good match for Wolffia arrhiza at around 250 µm.