On the other hand they agree with the Ganoids in having, in addition to the ordinary two divisions of the fish-heart, a third contractile chamber. This bulbus arteriosus is very different from the Bulbus aorta of the Teleosteous fishes, where it is simply a swelling of the walls of the aorta, not contractile, without valves in the interior, and separated from the heart by two valves opposite to each other. If this remarkable arrangement is deemed (and, I think, very justly) to be sufficient to separate the living Ganoids as a sub-class from the Teleostei, it is certainly significant enough to suggest the union of the Ganoids with the Plagiostomes. Moreover, this character is supported by two others of great importance, viz., the presence of a spiral valve in the intestine, which is found in a more or less developed state in all Ganoids, Sharks, and Rays, but is entirely absent, even in a rudimentary condition, in the Teleostei, and by the optic nerves being placed side by side, and not decussating, as is the case in all our ordinary fishes. Of the characters connecting these fishes I will refer to one other, as it has been described above, namely, that the fore and hind limbs of the Plagiostomes are also paddles supported by a cartilaginous structure, as in the Dipnoi.

The evidence in favour of a union of Ganoids with Plagiostomes is rendered complete by the Chimæras, which hold a surprisingly intermediate position. They are Sharks in external appearance and with regard to the structure of their organs of propagation; they are provided with the same copulatory organs, and their ova are large, enclosed in a horny case, and provided with adhesive appendages. Many species of Sharks, when in a very young state, are provided with a double dorsal series of spines (permanent in certain Rays), which are lost with age; and this most remarkable developmental character occurs likewise in young Chimeras. On the other hand, there is only one external gill-opening on each side, as, for instance, in Ceratodus, which, on the other hand, shows the first step towards a coalescence of the gills with the walls of the gill-cavity. The skeleton is notochordal, and the palatal and maxillary apparatus coalesce with the skull, as in Dipnoi, which is not the case in any Plagiostome; likewise the dentition approaches that of Ceratodus. Finally, Sir P. Egerton has drawn attention to the most important fact that the dorsal spine is articulated to a neural apophysis, and not merely implanted in the soft parts and immovable, as in Sharks.

Thus, then, the union of these fishes in one sub-class appears to me fully justified, as far as the living forms are concerned; but, as is implied by the name Palaichthyes, which I have proposed for this sub-class, it is intended to comprise also a great variety of forms from the Paleozoic Era, in fact, the predecessors of the Teleostean fish-fauna of the present period. I am aware of the objections that may be urged. First, it may appear to some to be an improper proceeding to unite in the same sub-class fishes of so different an appearance as a Shark and Lepidosiren, or as an Amia and a Pteraspis; but let them consider what a comprehensive category a sub-class necessarily is that the diversity between the fishes just named is not greater than that existing between a Sun-fish (Orthagoriscus) and an eel, or between a viviparous Embiotoca and a Loricaria, forms admitted by every ichthyologist of the present day as members of the same subclass, that of Teleosteans. In fact the Palæichthyes are composed of a similar series of modifications as the Teleosteans, some of the members of one sub-class exhibit ing marked analogies with those of the other, in the same manner as is the case with Placentalia and Implacentalia among mammals. To mix up ganoid-looking Teleosteans, like the Siluroids, with Ganoids, is as little in accordance with the advanced state of our ichthyological knowledge, as the union of Salamandra with Lacerta would be. Secondly, other naturalists may consider it very hazardous to establish a division, of which the majority of members are extinct

and known from remains of the hard parts only, and to characterise it by peculiarities of the soft parts. But why should we not make use of zoological evidence for the completion of the imperfect paleontological record, with the same benefit to science as in other cases, since not a few zoological problems have been, or can only be, solved by reasoning founded on palæontological facts? If, in the determination of affinities, we were to limit ourselves only to the consideration of those parts which have been preserved in the process of fossilisation, we could never expect any other result but the creation of most artificial assemblages of forms, although the characters of some natural families, or even orders, might be partly recognised.

*

On the one hand, we know that all the Teleosteous fishes, that is, the types which are predominant in the present and next preceding epochs, and which were but sparingly (Coccosteus?) represented in the Paleozoic, if they existed at all, agree, in spite of all other differences, with one another in possessing a two-chambered heart, with a rigid bulbus aorta and decussating optic nerves, and in never exhibiting a trace of a spiral valve in the intestine. On the other hand, we find that the few ichthyic types which have survived from the Paleozoic Era into our period, and those of which no immediate representative is known in that Era, but which approach that Amphibian fish-type by unmistakeable characters, agree, in spite of all other differences, in having a three-chambered heart, non-decussating optic nerves, and a spiral valve in the intestine. These are facts; and it seems to be a fair conclusion that the members of the Paleozoic fish-fauna had essentially the same organisation of those soft parts as their surviving representatives.

In conclusion I may shortly pass in review the living Palaichthyes, especially in regard to their distribution over the globe.

1. Of the order Plagiostomata or Marine Palaichthyes, 140 species of Sharks, distributed among 39 genera, are known, and 150 species of Rays, belonging to 25 genera. They inhabit nearly all the seas of the globe, decreasing in number from the tropics towards the poles. Only very few enter or live in freshwater.

2. The order Holocephala contains only four species, viz., three Chimæras and one Callocephalus ; they are restricted to the seas of the temperate zones of both hemispheres, and are absent between the tropics.

3. The order Ganoidei or Freshwater Palaichthyes is represented by one species of Amia, from North America; three species of Lepidosteus, from the same region, but extending southwards into Central America and Cuba; two species of Polypterus (Calamoichthys) from the tropical parts of Africa; two species of Polyodon, from the Mississippi and the Yantsekiang; about twenty-five Sturgeons, from the temperate and sub-arctic regions of the Northern Hemisphere; two species of Ceratodus from tropical Australia; one species of Lepidosiren from the Amazon river, and one of Protopterus from tropical Africa. Although the majority of the Sturgeons pass a part of the year in the sea, they must be regarded as freshwater fishes like the migratory Salmones, because they deposit their spawn in the rivers, where they also pass the first period of their growth; some species never enter the sea at any period of their life, and none are known to propagate in the sea.

The total number of fishes known at present being about 9,000, the Palaichthyes form only 36 per cent. of that number. But from the extent of the regions hitherto ichthyologically unexplored, and from the numerous additions annually made to the list of known forms, I do not believe that we are acquainted with much more than onetenth of the species of fishes actually existing.

ALBERT GÜNTHER

*From these considerations Amphioxus and the Marsipobranchii are excluded, the former being evidently the type of a distinct sub-class.

METEOROLOGY IN AMERICA * III. SELF-REGISTERING INSTRUMENTS

INVALUABLE as is the ordinary barometer, the most valuable instruments are those which are automatic, or self-registering. Prominent among those used in America are the Self-recording Barometer and Meteorograph invented by Prof. G. W. Hough, Superintendent of the Dudley Observatory. Lord Rosse's telescope has not done more for astronomy than will the self-registering barometer do for meteorology.

The diagram, Fig. 7, will illustrate the method of registering the height of the barometer and thermometer on a single sheet, by the use of one set of mechanism in these simple yet complete and consummate contrivances. Let D be a drum six inches in diameter and seven

inches in height, covered with a sheet of ruled paper. This drum is presumed to revolve at any convenient rate, say one inch per day. Let L be an iron or brass bar twenty-four inches in length, mounted on an axis passing through the point c. Let P be a steel pen attached to the end of the lever projecting over the centre of the drum. Let P' and P" be platinum wires attached to the lever at three inches on either side of the axis c. The wire P' is over the shorter leg of a siphon barometer, and the wire P" passes into the end of an open mercury thermometer. Now if the lever L be elevated at the end over the drum, the wire P' will touch the top of a float resting in the shorter leg of the siphon barometer If then a battery, B, and electro-magnet, E, be arranged as in the diagram, when contact is made with the float, a current of electricity will pass through the circuit, and the electro-magnet E is operated. If then, when the circuit is completed, a

blow be struck on the pen P, by means of the electromagnet, or a hammer unlocked by it, the dot on the drum sheet will indicate the height of the barometer at that time. It is obvious that as often as the lever is elevated a

record will be made. For the barometer an hourly record

will be found to be sufficient.

If the lever L is rigid and firmly mounted, the mere measurement of height by means of electrical contact can be carried to almost any degree of precision.

It was found from numerous experiments made some years since, that the magnetic circuit is not completed for a distance of one ten-thousandth of an inch. Therefore, whatever source of error there may be in the results recorded by this method is due to the barometer itself. In

We are again indebted to Harper's New Monthly Magazine for the continuation of the article by Prof. Maury, and the woodcuts which we reproduce this week.

practice, from records extending over nearly one year, it is found that the results are inside the errors of reading from the drum sheet.

A long experience has led to the conclusion that this degree of precision is sufficient for the investigation of barometric changes, and is but little outside the limit of error from reading a standard barometer.

An examination of the diagram will also show at a glance how the height of the thermometer is recorded. It should, however, previously be stated that the thermometer is a little larger than those in ordinary use, and has a platinum wire, a, cemented in the bulb, communicating with the mercury in the inside.

The following is a general description of a machine constructed for the Signal Service at the request of the chief signal officer.

It registers hourly the barometer and wet and dry bulb

rest during a period of fifteen minutes, the time required for elevating the hammers H and H'. As soon as this is accomplished, the lever begins to rise slowly by means of the double snail on the hour shaft, the time required for traversing the drum being about fifteen minutes. When the position of the lever is such that the carriage in the rear of the clock touches the float in the shorter leg of the siphon, an electric current is established through the magnet, F, which unlocks the hammer H, causing the pen G to make a record on the drum sheet. After the lever has reached the top of the drum, it remains at rest fifteen minutes while the hammers are being raised, when it is gradually depressed. So soon as the platinum wires attached to the carriage over the thermometers-touch the

surface of the mercury in the thermometer tubes, electric currents are established through the magnets F and ], simultaneously or successively unlocking the hammers, and, as the case may be, making records as before. A complete double motion of the lever requires one hour. During this time the barometer and wet and dry bulb thermometers have each been recorded once. The records of the barometer and thermometers differ in time about half an hour. The wet and dry bulb thermometers are recorded within about one minute of each other, depending on the difference between them.

One of the most marked and wonderful features of the invention of Prof. Hough is that it prints its own records. And this is done by a single screw, which rises or falls

with the mercury in the barometer. This screw carries a pencil, which traces upon a revolving cylinder or roll of paper a line showing the minutest movements of the column of mercury for every minute in twenty-four hours. This same screw also gives motion to a series of wheels which carry types, by which, at the end of every hour, the height of the column of mercury is printed on a slip of paper to the accuracy of the thousandth part of an inch!

One of the most beautiful and simple contrivances used is a Wild's Self-registering Barometer, of which we give a cut one quarter the actual size. It scarcely needs explanation, except to say that the tube A is suspended in a

cistern of mercury, represented on the left of Fig. 9. As the atmospheric pressure changes, the level of the mercury changes in the cistern, and the tube A rises or falls as the atmospheric pressure increases or diminishes. The weight of this tube as it floats in the mercury, and also that of the arm I, which supports it at G, is exactly balanced by the arm 11, to which is attached a sliding weight, III, adjustable by a small thumb-screw. K is a steel crayonholder fixed to the balance III, and to which is fixed a crayon, c, whose point in seen in B to impinge upon a When the atmospheric pressure is increased, the tube 4 sheet of paper, . This sheet s moved by clock-work. is forced to rise a little out of the mercury in which it

B

bats, and as it rises at G, the arm / is elevated. The crayon older, being fixed on the balance at the fulcrum, f, by two tle screws, swings a little to the left, and the crayon hich it carries with it makes a mark on the paper beeath it, which mark indicates the rise of the barometer, r the increase of atmospheric pressure. If the pressure ecreases, the pencil, of course, moves in the opposite irection, and shows the barometric fall. The roll of aper on which the record is made by this automatic nstrument is divided into rectangular parts, each one of which exhibits the atmospheric variations for twenty-four ours. At the end of every day this part of the roll is detached and put by to be bound up in book form in the records of the office in which the instrument is kept.

nothing is wanting to give reliability to the published results and the "probabilities" issued from his officers. A self-registering barometer, as well as other instruments

THERMOMETERS.

FIG. 10.-DRAPER'S PHOTOGRAPHIC REGISTER OF BAROMETER AND THERMOMETER AT NEW YORK, APRIL 28, 1871 The upper line of the Thermometer is the Dry Bulb, the lower line is the Wet Bulb

FIG. 11.-PHOTOGRAPH OF A STORM.-(Print from Photographic Register from Noon, December 11, 1870, half-inch per hour.) Two Thermometer and One Barometric Curve

The roll of paper is on a reel, n, passing between two rollers, g and k, as seen in B (Fig. 9).

By these perfectly simple devices, instead of obtaining only three daily recorded observations, the observer at every station gets a continuous and perpetual record for every second in the day. That is to say, instead of getting, as by the common barometer (observed three times a day), observations for three seconds in twenty-four hours, he gets them for as many seconds as there are in twenty-four hours, or 86,400. Thus it follows that the value of the self-registering barometer, as compared with the ordinary one, is as 86,400 to 3!

The marvellous accuracy and exquisite nicety with which all the observations forwarded to General Myer by the observers are marked ought to assure the public that

of equal sensitiveness, will be used by all the observersergeants. It is scarcely possible for this invaluable instrument to suffer derangement or to get out of order. A third most beautiful and sensitive self-registering