EVIDENCE FOR A CONDUCTING STRAND IN EARLY SILURIAN (LLANDOVERIAN) PLANTS: IMPLICAnONS FOR THE EVOLUTION OF THE LAND PLANTS

-Macerations of fragmented plant compressions of Silurian (Llandoverian) age yield sheets of organic material bearing ridges and depressions (interpreted as corresponding to surficial dimensions of cells), fragments of smooth walled tubular elements, and fragments of tubular elements with differentially thickened walls ("banded tubes"). A fragment of tissue (1.2 mOl long), consisting of smooth walled tubes (17 ± 6.9 ""'01 in diameter), surrounding 2-3 larger (18.8 ± 2.1 ""'01 in diameter) banded tubes, wasisolated from an irregularly shaped, relatively large (1 x 3 mOl) compression. Comparisons between (I) fragments of tubular cell types, not organized into strands, isolated from 122 compressions, and (2) dispersed tubular cell types previously reported from the Massanutten Formation and other Silurian formations (Tuscarora and Clinton), reveal no significant morphologic differences. Comparisons between the organization of smooth-walled and banded tubular cell types found in the tissue strand and the or~anization of cell types in nematophytic plants (Nematothallus, Nematoplexus, Prototaxites) indicate a similarity in construction (tubular) hut a lack of correspondence in organization. The strand of tissue is interpreted as representing part of the internal anatomy of a nonvascular land plant of unknown taxonomic affinity. On the ba-,is of analogy with present-day embryophytes, the strand of tubular cell types is inferred to have functioned as a conductive tissue. The significance of "banded tubes" in Silurian strata is discussed, and it is concluded that, until more is known about the anatomy, morphology, and biochemistry of the parent plant(s), the habitat and systematic affinity of these organisms are conjectural. Karl 1. Niklas and Vassiliki Smocovitis Section of Plant Biology, Cornell University, Ithaca, New York 14853 Accepted: May 3 I, 1983 Introduction analogous in function to supportive or conduct­ Previous studies of Llandoverian and Wen­ ing plant cell types. lockian compression fossils from North America This paper describes additional plant have recognized three categories of microfossils: compressions collected from the localities pre­ (1) tubular elements of variable length having viously reported by Pratt et al. (1978) in the smooth and banded (=differentially thickened) lower Massanutten Sandstone, Virginia. Among walls, (2) membranous sheets of organic mate­ the various cell and tissue fragments recovered rial bearing irregularly shaped depressions, and from 122 plant fragments, a single strand of tis­ (3) alete and trilete spores, and tetrads (Gray sue was isolated consisting of smooth-walled tu­ and Boucot 1977; Pratt et aI. 1978; Strother and bular elements surrounding a small number of Traverse 1979). Llandoverian microfossils from banded tubular cell types. This fragment of tis­ the Massanutten Formation of Virginia are sue, interpreted to be part of the internal anat­ thought to be the fragmented remains of a non­ omy of an as yet unspecified macrophyte, is the vascular macrophyte of unknown taxonomic af­ largest fragment thus far reported for the Mas­ finity based on an inferred fluvial depositional sanutten localities, and provides evidence for a environment, the presence of spores, and chem­ discrete organization of smooth and banded ical composition (Pratt et aI. 1978; Niklas and tubes. In addition to the new morphologic in­ Pratt 1980). Owing to poor preservation, the formation, stable isotope (6C ) analyses of du­ general morphology and internal organization plicate fossils are presented. In conjunction with of these plants were previously unknown. How­ previously reported organic chemical profiles, er I Pratt et aI. (1978) have speculated that the these data provide some insight into the bio­ th and banded tubes represent fragments chemical composition of the Massanutten plant bular-filamentous tissue (nematophytic), remains. The additional morphologic and bio­ , e Paleontological Society. All rights reserved. 0094.8373/8310902--o00S/$I.00 127 EVOLUTION OF LAND PLANTS chemical data presented here and elsewhere (d. Niklas 1982) are relevant to the biologic inter­ pretation of the Massanutten fossils and provide a context for evaluating current concepts on the first occurrence of the land plants and the sub­ sequent appearance of tracheophytes (Banks 1975; Gray and Boucot 1977, 1979). Materials and Methods Samples of fossiliferous siltstone were collect­ ed from horizons designated by Pratt et al. (1978) as 8-9 in the lower Massanutten Sandstone along route 678 at the gap of Passage Creek, 6 km southeast of Strasburg, Virginia. The age of the Massanutten compression fossils is reported as most probably Llandoverian A or B (Pratt et al. 1978). Two shale partings, each consisting of part and counterpart, measuring 16 x 8 em and 17 x 9 em were examined, and revealed 122 compres­ sion fossils on their bedding surface (roughly 281 cm2). The maximum lateral dimension of each compression was measured with a Helios mi­ crocaliper (Modern Tools Corp., Woodside, N. Y.) and a histogram of size classes construct­ ed. Additional shale partings with compressions were reserved for stable isotope analyses. Spec­ imens were photographed with reflected light, and representative specimens were removed from the larger partings for SEM by means of a dental drill cutting rotor. Macerations of compressions, grouped according to size class (0.26-0.50,0.51-0.75, ... ,2.51-2.75 mm, cE. Fig. 1), were obtained by constructing paraffin wells around specimens and treating them with HF acid for 30 min to 48 h. Macerations were neutralized with washings of distilled water, and organic debris removed by pipette for light and scanning microscopy. Stable isotope (OC ) analyses of compression fossils were performed by Krueger Enterprises, Inc., Geochron Laboratories Division (24 Blackstone Street, Cambridge, Mass.), and are based on samples designated CR-16669 and CR­ 16670 by that facility. Statistical comparisons among the morpho­ logic features of tubular cell types are based on standard 2-sample t-test, corrected by a Bon­ ferroni inequality to df 0.05/N (where N = the number of comparisons and df = degrees of freedom) (cf. Snedecor and Cochran 1980). 30 SKEWNESS" fbt = 0 614* 25 24 25 KURTOSIS'" KURTOSIS = bz = 254 t 0 =0.19* "z w n = 122 i = 1.32 ! 20 19 v =.!051 u w "­ 16 "­ o 15 13