mouth of the spire, however, rapidly opening out by successive additions, so as to enclose the nucleus'; after which all succeeding additions were complete rings, so that the cyclical plan came to be completely established.-This I designated as the 'simple' type of Orbitolite structure.

I was further able to show that these two typical forms were connected by a gradational series of connecting links; the formation of disks of the complex' type often commencing on a plan resembling that of the simple;' and the change from the latter to the former taking place, not at any fixed epoch of growth, but after a variable number of rings had been formed, sometimes abruptly, sometimes more gradually, in the manner to be presently detailed. And I also found that the inner rings of even the largest 'complex' disks, if their early growth had taken place on the 'simple' type, were not complete, but showed a tendency to one-sided and therefore spiral growth, like that seen in Fig. VI., 3.

Reflecting on the relations of these highly specialised Foraminiferal types to the simpler forms of the Milioline group, to which (in virtue of the 'porcellanous' character of their shells) I referred them, I ventured to construct a hypothetical pedigree; tracing their descent (Fig. V.) from the particle of protoplasm that forms the spheroidal chamber in which every Foraminiferal shell begins, first to an open undivided spiral (1); then to a type in which the spire is constricted at intervals (2); then to a type in which it is completely divided into chambers by transverse partitions (3); then to a type in which the spirally arranged chambers are divided by longitudinal partitions into chamberlets (4); then to the 'simple' type of Orbitolites, in which the spiral plan of growth gives place to the cyclical (5); then to an 'intermediate' type, in which the original spiral almost disappears (6); and finally to the 'complex' type, in which the plan is cyclical from the beginning (7).

This hypothetical pedigree has found its complete confirmation in a deep-sea Orbitolite of extraordinary delicacy

and beauty, which was brought up in the Porcupine Expedition of 1869. For this little disk, about the size of a fourpenny piece, while for the most part truly cyclical,

Diagram illustrating the Pedigree of the Complex type of Orbitolite.

1. Simple undivided Spire of Cornuspira.

2. Partially interrupted Spire of Spiroloculina.

3. Spire of Peneroplis, divided by partitions into chambers.

4. Spire of Orbiculina, its chambers divided into rows of chamberlets.

5. Disk of 'simple' Orbitolite, showing first-formed spire, surrounded by concentric rings.

7. Central portion of Disk of complex' Orbitolite, in which the chambered nucleus alone shows an abbreviated spire, the very first row of chamberlets forming a complete ring.

has a long succession of inner chamberlets arranged upon the spiral plan, as in Orbiculina; these, again, arise from expanded but undivided chambers, like those of a Peneroplis ; and these chambers are the continuation of a spiral tube, with occasional constrictions, resembling that of a Spiroloculina, coiling continuously round a primordial chamber, as in Cornuspira. Thus, in this interesting organism we find permanently represented the whole developmental history of the 'simple' type of Orbitolite from the primordial jellyspeck. The large Challenger collection of Orbitolites, made on the Fiji reef, has furnished me with the means of still more completely working out the transition from the 'simple' to the 'complex' type; a distinctly intermediate type there presenting itself in great abundance. This, which I term the duplex' type (Fig. VI., 1), resembles the 'simple' in having its annular series of chamberlets disposed in a single plane, and in the connection of the chamberlets of each ring by a single annular canal; but differs in having its successive rings connected by a double series of radial passages, which issue on the edge of the disk (Fig. VI., 2) as marginal pores. The columnar sub-segments, a a', b b', of each ring are strung, as it were, on the annular cord, c c'; and this sends off an upper and a lower series of stolon-processes, d d, d'd, which pass into the upper and lower halves of the sub-segments of the next ring.-The plan of growth in the first-formed portion, shown in Fig. VI., 3, is singularly intermediate between that of the 'simple' and that of the complex' type. The regular spire of the former is now reduced to the single turn made by the 'circumambient segment', b b, round the 'primordial segment', a; but a partial continuance of the same plan is shown in the incompleteness of the first two or three rings of sub-segments; these being budded-forth from only half of the circumambient segment,' instead of from its whole periphery, as in the typical 'complex' Orbitolite (Fig. II.). Yet even in large disks, whose later growth is characteristically complex,' the nucleus and earlier rings

are often formed on the duplex' plan, which passes into the 'complex' in the manner to be now described.

Believing, with Sir James Paget, that "the highest laws of Biological Science are expressed in their simplest terms in the lives of the lowest orders of creation," I shall now ask

1. Disk of Duplex type of Orbitolite.

2. Edge of Disk, showing double row of marginal pores.

3. Central portion of Sarcode body :-a, primordial segment; b, circumambient segment, budding off a half-ring of sub-segments, from which complete rings are afterwards formed.

4. Portion of the Sarcodic body of one ring; a a' and b b', the two halves of the columnar sub-segments in connection with c c', the annular cord; from this are given off the pairs of stolon-processes d d', d d', which connect it with the sub-segments of the next annulus.

you to follow me through a detailed examination of the transition from one type to the other; as shown in Fig. VII., which represents a vertical section, taken in a radial direction, of one of those large complex' disks whose life was commenced on the plan of the 'simple.' The firstformed series of chamberlets (m, m1, m2, m3, m1) exactly correspond with those of the 'simple' type (Fig. IV., 4), constituting but a single plane; those of each series being connected together by a single continuous annular gallery (shown in cross-section at ac, ac), while those of each series are connected with those of the next by single radial passages (r, r, r), which, as each annulus was formed, would open at its outer edge as a single row of marginal pores. But these are surrounded by rings (d, ď, ď2) in which, while the annular canal is still single, two radial passages (r) go off from it obliquely, one into the upper and the other into the lower portion of each chamberlet of the next annular series, those of the last-formed annulus showing themselves at its edge as a double row of marginal pores. From this duplex' type, the first advance towards the 'complex' is shown at e, e', in the splitting, so to speak, of each annular canal into two (ac, ac'), and the interposition of a columnar cavity (m, m) between its two halves. Now, in the inner (or earlier-formed) of the annuli which show this complication (e, e1), the two series of chamberlets (s s, s' s') which lie between the two annular canals and the two surfaces of the disk, are continuous with the intermediate columnar chamberlets, and bear the same relation to their respective annular canals as in the 'duplex' type, each being connected with one canal only; and this stage of differentiation characterises the Orbitolites of the French Tertiaries, which seem to have attained their full growth without any advance upon it. But in the large Orbitolite disks of Australia and Fiji, I find this simpler arrangement giving place to a more complicated one (f, ƒ1‚ƒ3‚ƒ3); the chamberlets of the two superficial layers being separated from those of the intermediate layer, and being so shifted in position, that each annular series lies over the