Lampbrush chromosomes isolated in saline from oocyte nuclei of newts have been examined by means of a phase-contrast microscope with inverted optical train. The newts which have been studied-four subspecies of Triturus cristatus; cristatus, carnifex, danubialis and karelinii-have diploid complements of twenty-four chromosomes in somatic cells and twelve lampbrush bivalents in each oocyte nucleus. Though the structural and functional organization common to all lampbrush chromosomes is discussed in some detail, this paper is mainly concerned with descriptions of those morphological characters which serve to identify each particular chromosome. The axis of a lampbrush chromosome consists of a series of chromomeres connected to one another by short, thin, extensible and elastic filaments. Loops and other objects of a wide range of morphologies are attached laterally to the chromomeres, loops occurring in pairs. When lampbrush chromosomes are broken mechanically, each break occurs transversely across a chromomere, and separates the loop insertions in this chromomere in such a way that a pair of straightened-out lateral loops bridge the gap in the chromosome axis. A thin fibre forms an axis to each lateral loop, and around this axis lies 'matrix'. Chromomeres contain deoxyribonucleic acid (DNA), and the interchromomeric filaments and loop axes consist of DNA fibres. These DNA components of lampbrush chromosomes are envisaged as the persistent genetic material. Matrix, which consists of ribonucleoprotein (RNP), is assumed to be gene product. Although many lateral loops have similar morphologies, their matrices consisting of very fine fibres projecting radially from loop axis, certain loops are identifiable by peculiarities of matrix texture and quantity. These readily distinguishable structures have been used as `landmarks' to map the chromosomes, and together with relative axial chromosome lengths they serve for chromosome recognition. The newt chromosomes have been designated I, II, III<latex>$\ldots$</latex>XII in diminishing order of relative length. Fusion is a property of some types of loop matrix, and it can occur within single loops, between sister loops or between homologous loops. It can also occur between non-homologous loops, though such fusions are not haphazard and involve loops of similar matrix texture. Wherever loop form is obliterated by matrix fusion the underlying loop structure can be demonstrated if the loop matrix is partially dissolved in dilute saline. Certain objects attached to newt lampbrush chromosomes are not organized about a loop basis. Thus telomeres and `axial granules' each consist of a crescent of DNA to which a spherical or roughly spherical mass of RNP is closely applied. All conceivable combinations of homologous and non-homologous fusions between these structures can occur, many such fusions involving more than two comparable objects. Chromosomes V and VIII carry `spheres' attached directly to certain chromomeres, and many of the conceivable types of fusion between these spheres have also been observed. Non-homologous fusions between comparable objects on one and the same chromosome are described as `reflected' fusions in contradistinction to non-homologous fusions between different chromosomes. Reflected fusions between certain axial granules on chromosomes II, III, IV and VI occur so frequently that they are useful diagnostic characters. Homologous lampbrush chromosomes are joined to one another by chiasmata and, as just mentioned, they may also be associated by gene product fusions. All chiasmate associations are junctions between chromosome axes; chiasmata do not occur within the lengths of lateral loops. In all four subspecies chiasmata tend to occur more frequently in regions close to the centromeres than elsewhere, and in subspecies carnifex analysis of this chiasma distribution enabled the centromeres to be identified. In carnifex the centromeres are spherical objects, without lateral loops, which lie in the chromosome axes and are slightly larger than the majority of chromomeres. In fixed preparations, in contradistinction to axial granules, they stain Feulgen-positive throughout. Cristatus centromeres are similar but smaller, danubialis centromeres smaller still. In these three subspecies homologous chromosomes are never associated precisely at their centromeres, and this latter feature was indeed mainly responsible for their identification. In subspecies karelinii the centric regions are much more conspicuous, the centromere granule being flanked by thick, Feulgen-positive portions of chromosome axis bare of lateral loops. As karelinii oocytes increase in size these thick portions of axis lengthen by the incorporation of neighbouring chromomeres, whose lateral loops meanwhile regress. A comparable process of lateral loop regression and amalgamation of chromomeres occurs throughout lampbrush chromosomes as oocytes reach maturity, and with regard to this phenomenon the centric regions of karelinii are thus precocious. Unlike the centromeres of the other three subspecies, homologous centromere granules of karelinii are often fused together, and non-homologous and threefold centric fusions have also been observed in this subspecies. Despite these outstanding differences the centromeres of carnifex and cristatus have without doubt been correctly identified, and they are located at substantially the same places on the chromosomes as those of karelinii. Although the chromosome complements of the four subspecies show overall resemblances, other characteristics apart from the centromeres serve to distinguish between them; and these are detailed. It is significant that loops of certain particular morphologies are present in one or other of the subspecies, but not in the remainder. As a general rule homologous chromosome sites bear lateral objects of comparable size and texture. However, in all four subspecies the two largest chromosomes which form bivalent I are never associated by chiasmata within regions extending over about half their length, and in these regions the lateral loop patterns do not correspond. Moreover, in certain female newts at specified homologous sites on other bivalents the lateral loops may regularly be of dissimilar morphology. These differences are conserved throughout life. Several subtle distinctions of this kind are detailed, but there are also more striking individual-specific characters. Thus in carnifex there are females with giant loops present on one homologue only of bivalent X, other females with giant loops on both homologues, and yet other females without giant loops on either. Within a population the proportions of these three types of female accord with Hardy-Weinberg expectations, and such individual characteristics are claimed to represent, with respect to gene products, allelic differences at particular genetic loci. Liberation of synthesized products from the majority of lateral loops has not been observed, but from many of the landmark loops where gene products accumulate in massive quantities it can be inferred that aggregates of these materials are shed from time to time into the nuclear sap. Detached gene products, whose sites of origin have been identified, diminish in size after shedding and they presumably augment the nuclear sap. The trivial variation in morphology of particular lateral loops from one oocyte to another, taken from a single female, is claimed to be due to varying balance between the rates of gene product synthesis and dispersal existing at the time of dissection. The origin and functional significance of the peripheral `nucleoli', which are so characteristic of amphibian oocyte nuclei, are still uncertain. Lateral loops are characteristically asymmetric. One of the two portions of loop axis leading from a chromomere is always relatively bare of matrix. Along the lengths of the great majority of lateral loops there is an even transition in matrix quantity and/or texture, but loops of highly irregular outline also regularly show thinner and thicker insertions in chromomeres. Wherever axial breaks occur the polarity of loop asymmetry with respect to the whole chromosome can be determined. Sister loops always show the same polarity, and at those places where axial breaks have been repeatedly observed loop asymmetry consistently displays the same polarity. This fundamental feature of the genetic material is claimed to depend on a polarized mechanism of loop extension from and retraction to its parent chromomere, one end of the loop having therefore been engaged for less time than other parts in synthesis of matrix. If the main and subsidiary arguments supporting this claim are valid, each loop axis must consist of a series of repeats of identical genetic information. To reconcile this theory with the dimensions of mutational sites, so far as these are known, in other organisms, it is suggested that each chromomere consists of a `master copy' of the genetic code and a store of incompletely specified DNA. Further specification of DNA from this store occurs whilst the loop axis extends, and it is this secondarily coded material which acts as the template for synthesis of specific RNP.