Sporobolus spikelet structure, illustrated with S. compositus var. macer.

The spikelet consists, from the base up, of 2 glumes (spikelet bracts: lower [g1] and upper [g2]), and a floret with two bracts (lemma [L] & palea [P]). The floral stigma & anthers may be visible at anthesis; the ovary develops into a grain [G]. When mature the spikelet disarticulates at the glumes (i.e., the floret falls off as a unit, leaving the glumes behind).

by Bob Harms 

The following table shows the great variation in lengths/widths for a single population of S. clandestinus, even in August, when most panicles/spikelets are immature. Although these dimensions do not distinguish the two most generally confused taxa, S. clandestinus & S. compositus var. compositus, spikelet length clearly separates (1) var. drummondii from other varieties of S. compositus; (2) S. neglectus from S. vaginiflorus; and (3) additionally, within S. vaginiflorus relative glume/floret length has been used to distinguish two varieties.

1. S. compositus var. drummondii. My own 2002 survey of TEX/LL holdings found that all 30 collections with glabrous lemmas with a spikelet length less than 4.2 mm (i.e., 3.2—4.1 mm) were S. compositus var. drummondii; the 2 other varieties had longer spikelets (4.5—6.4). All three outside experts who examined nearly all of these same collections also determined them as var. drummondii (with one exception, a plant deemed to be S. vaginiflorus by one expert). And no one considered a plant with larger spikelets and glabrous lemmas (13 collections) to be var. dummondii.

   Strangely, current keys and descriptions don't recognize this varietal differentia. I also note that Jones et al. merge this variety with var. compositus.

2. S. neglectus. Most recent treatments use spikelet length to distinguish S. neglectus from S. vaginiflorus; e.g., Peterson et al: "spikelets 1.6-3 mm long" vs. "spikelets 2.3-6 mm long."
3. S. vaginiflorus glume/floret length. Treatments that recognize the two varieties, var. ozarkanus & var. vaginiflorus, use glume length—floret length as differentia; e.g., Peterson et al.
   • Sheath bases sparsely hairy; glumes usually longer than the florets; lemmas always faintly 3-veined … var. ozarkanus
   • Sheath bases usually glabrous; glumes usually shorter than the florets; lemmas usually 1-veined … var. vaginiflorus

   But note the hedge wording (i.e., "usually") in their key. In surveying the 49 TEX/LL S. vaginiflorus collections in 2002 I found that where a sheet had notations such as 'glumes (mostly) longer than floret' they were determined to be var. ozarkanus. But this characteristic was not consistently applied — even specimens with 3–nerved lemmas and glumes longer than florets were determined to be var. vaginiflorus. Specimens with glumes both longer and shorter than the floret were determined as var. ozarkanus. In my view the distribution of these characters, as plotted in the map below, does not support recognition of variety taxa.

   Lemma venation and glume/floret lengths for S. vaginiflorus forms in Central Texas counties (as noted in my 2002 survey).
   Symbols do not represent locations below the county level.

Sporobolus clandestinus

Spikelets at anthesis; September; Hays Co.

Spikelets, florets; October; Hays Co.

Sporobolus compositus var. compositus

Spikelets; September; Parker Co.

Spikelets; October; Colorado Co.

Sporobolus compositus var. macer

Spikelets at anthesis; September

Spikelets; November

Sporobolus compositus var. drummondii

Spikelets at anthesis; September; Hays Co.

Spikelets; October; Hays Co.

Sporobolus vaginiflorus var. vaginiflorus
²Variety determined following the FNA key.

Anthesis; September — Spiklets, florets; November;   Hays Co.

Sporobolus vaginiflorus var. ozarkanus^2
2Variety determined following the FNA key.

    
Terminal spiklets, glumes, florets; November

Sporobolus neglectus

TEX/LL herberia have a total of 5 S. neglectus collections, and only one of these (Riggins 396, Grayson Co., on the Oklahoma border) has numerous spikelets, shown below. 2 were from Grayson Co. 7 other collections (of S. vaginiflorus) from Travis Co. have incorrectly been determined as S. neglectus. The distinction is not subtle. Note that spikelets are significantly shorter, less than 3 mm, used in the FNA key to distinguish S. neglectus from S. vaginiflorus.

Spiklet from sheath; Grayson Co.

Spiklets extending from sheath; Grayson Co.