Smith et al.: Morphological Evidence
473MATERIALS AND METHODS Ammocoete larvae of the lamprey P . marinus were provided by Mr. E. L. King, Acting Investigation Chief, Bureau of Commercial Fisheries, Hammond Bay Biological Station, Millersburgh, Michigan. Larvae were maintained in shallow tanks of aerated water at 4" C. Details of fixation of the central nerve cord, and subsequent processing and examination have been described previo u~l y .~~ Note that cross-bridges between axoplasmic microtubules and membrane-limited structures including mitochondria are most evident after heavy staining, up to one hour in saturated 50% ethanolic uranyl acetate followed by 10-20 minutes in lead citrate. As will be shown, material illustrated in FIGURES 26 and 27 was treated with vinblastin (0.01 M/50 1. 1) injected into the perineural space exposed by removal of a short segment of the caudal region and then sealed by a suture clip. For quantitative observations on the organization of the axoplasm and notably for mitochondria/microtubule counts, the instrument magnification was assessed from micrographs of a diffraction-grating replica photographed at the same instrument settings as for electron-image plates used for recording representative fields of transversely sectioned material used for preparation of counts. In the case of large (Muller) fibers, which in the 12-14 cm larvae employed reach a diameter of 45 microns, up to 20 plates were exposed and printed as a montage at a final magnification ( X 50,000) permitting adequate measurement with a magnifying 0.1 mm scale. Sections of small axons were photographed at higher instrument magnifications following similar recording of a standard grating replica. Sections were examined in a Philips EM 200 and EM 300.
Cornputer Analysis of Randomly Organized Axon ModelsCounts of microtubules and mitochondria in "small," "medium," and ''large'' diameter axons were made, as mentioned above, from electron micrographs. These groups were arbitrarily designated as axons 2-10, 15-20, and 30-45 p in diameter. The last group are the giant Muller interneurones illustrated and discussed in a previous account.' Counts were used to prepare models of axons of specified diameter, in which a strictly random, nonassociative "cytoplasmic architecture" is assumed. These models of the distribution of mitochondria and microtubules were prepared on a UNIVAC 1106-1 108 computer. The diameter of four actual axons ( 3 large and 1 medium) and the average diameter of 18 small axons were taken, together with the average cross-sectional diameter of a mitochondrion and a microtubule. Counts of mitochondria and microtubules were taken from these axons. The basic parameters are summarized in TABLE 1 (RESULTS).The model of each axon type, five in all (Ll, L,, L,, M and S ) , was prepared by drawing pairs of numbers uniformly distributed on the interval ( 0 , l ) by a pseudorandom number generator based on an optimized power residue method.'i These were taken as ( X , Y ) coordinates of the center of an object (mitochond...