Chapter CXLVII: Act 1875: , but, broadly speaking, they are divided into two (3)

1. Thermal Ammeters.--These instruments are also called hot-wire ammeters. In their simplest form they consist of a wire through which passes the current to be measured, some arrangement being provided for measuring the small expansion produced by the heat generated in the wire. This may consist simply in attaching one end of the wire to an index lever and the other to a fixed support, or the elongation of the wire may cause a rotation in a mirror from which a ray of light is reflected, and the movement of this ray over a scale will then provide, the necessary means of indication. It is found most convenient to make use of the sag of the wire produced when it is stretched between two fixed points (K1K2, fig. 1) and then heated. To render the elongation evident, another wire is attached to its centre S2, this last having a thread fixed to its middle of which the other end is twisted round the shaft of an index needle or in some way connected to it through a multiplying gear. The expansion of the working wire when it is heated will then increase or create a sag in it owing to its increase in

FIG. 1.--Diagram showing the arrangements of Hartmann and Braun's Hotwire Ammeter.

length, and this is multiplied and rendered evident by the movement of the index needle. In order that this may take place, the heated wire must be flexible and must therefore be a single fine wire or a bundle of fine wires. In ammeters for small currents it is customary to pass the whole current through the heating wire. In instruments for larger currents the main current passes through a metallic strip acting as a bye-pass or shunt, and to the ends of this shunt are attached the ends of the working wire. A known fraction of the current is then indicated and measured. This shunt is generally a strip of platinoid or constantin, and the working wire itself is of the same metal. There is therefore a certain ratio in which any current passing through the ammeter is divided between the shunt and the working wire.

Thermal ammeters recommend themselves for the following reasons:--(1) the same instrument can be used for continuous currents and for alternating currents of low frequency; (2) there is no temperature correction, (3) if used with alternating currents no correction is necessary for frequency, unless that frequency is very high. It is, however, requisite to make provision for the effect of changes in atmospheric temperature. This is done by mounting the working wire on a metal plate made of the same metal as the working wire itself; thus if the working wire is of platinoid it must be mounted on a platinoid bar, the supports which carry the ends of the working wire being insulated from this bar by being bushed with ivory or porcelain. Then no changes of external temperature can affect the sag of the wire, and the only thing which can alter its length relatively to the supporting bar is the passage of a current through it. Hot-wire ammeters are, however, liable to a shift of zero, and means are always provided by some adjusting screw for slightly altering the sag of the wire and so adjusting the index needle to the zero of the scale. Hot-wire ammeters are open to the following objections:--The scale divisions for equal increments of current are not equal in length, being generally much closer together in the lower parts of the scale. The reason is that the heat produced in a given time in a wire is proportional to the square of the strength of the current passing through it, and hence the rate at which the heat is produced in the wire, and therefore its temperature, increases much faster than the current itself increases. From this it follows that hot-wire ammeters are generally not capable of giving visible indications below a certain minimum current for each instrument. The instrument therefore does not begin to read from zero current, but from some higher limit which, generally speaking, is about one-tenth of the maximum, so that an ammeter reading up to 10 amperes will not give much visible indication below 1 ampere. On the other hand, hot-wire instruments are very ``dead-beat,'' that is to say, the needle does not move much for the small fluctuations in the current, and this quality is generally increased by affixing to the index needle a small copper plate which is made to move in a strong magnetic field (see fig. 2). Hot-wire instruments working on the sag principle can be used in any position if properly constructed, and are very portable. In the construction of such an instrument it is essential that the wire should be subjected to a process of preparation or ``ageing,'' which consists in passing through it a fairly strong current, at least the maximum that it will ever have to carry, and starting and stopping this current frequently. The wire ought to be so treated for many hours

FIG. 2.--Hot-wire Ammeter.

before it is placed in the instrument. It is also necessary to notice that shunt instruments cannot be used for high frequencies, as then the relative inductance of the shunt and wire becomes important and affects the ratio in which the current is divided, whereas for low frequency currents the inductance is unimportant. In constructing a hot-wire instrument for the measurement of high frequency currents it is necessary to make the working wire of a number of fine wires placed in parallel and slightly separated from one another, and to-pass the whole of the current to be measured through this strand.

In certain forms, hot-wire instruments are well adapted for the measurement of very small alternating currents. One useful form has been made as follows:--Two fine wires of diameter not greater than .001 in. are stretched parallel to one another and 2 or 3 mm. apart. At the middle of these parallel wires, which are preferably about 1 m. in length, rests a very light metallic bridge to which a mirror is attached, the mirror reflecting a ray of light from a lamp upon a screen. If a small alternating current is passed through one wire, it sags down, the mirror is tilted, and the spot of light on the screen is displaced. Changes of atmospheric temperature affect both wires equally and do not tilt the mirror. The instrument can be calibrated by a continuous current. Another form of hot-wire ammeter is a modification of the electric thermometer originally invented by Sir W. Snow Harris. It consists of a glass bulb, in which there is a loop of fine wire, and to the bulb is attached a U-tube in which there is some liquid. When a current is passed through the wire, continuous or alternating, it creates heat, which expands the air in the bulb and forces the liquid up one side of the U-tube to a certain position in which the rate of loss of heat by the air is equal to the rate at which it is gaining heat. The instrument can be calibrated by continuous currents and may then be used for high frequency alternating currents.

2. Electromagnetic Ammeters.--Another large class of ammeters depend for their action upon the fact that an electric current creates an electric field round its conductor, which varies in strength from point to point, but is otherwise proportional to the current. A small piece of iron placed in this field tends to move from weak to strong places in the field with a force depending on the strength of the field and the rate at which the field varies. In its simplest form an electromagnetic ammeter consists of a circular coil of wire in which is pivoted eccentrically an index needle carrying at its lower end a small mass of iron. The needle is balanced so that gravity compels it to take a certain position in which the fragment of iron occupies a position in the centre of the field of the coil where it is weakest. When a current is passed through the coil the iron tends to move nearer to the coil of the wire where the field is stronger and so displaces the index needle over the scale. Such an instrument is called a soft-iron gravity ammeter. Another type of similar instrument consists of a coil of wire having a fragment of iron wire suspended from one arm of an index needle near the mouth of a coil. When a current is passed through the wire forming the coil, the fragment of iron is drawn more into the aperture of the coil where the field is stronger and so displaces an index needle over a scale. In the construction of this soft-iron instrument it is essential that the fragment of iron should be as small and as well annealed as possible and not touched with tools after annealing; also it should be preferably not too elongated in shape so that it may not acquire permanent magnetization but that its magnetic condition may follow the changes of the current in the coil. If these conditions are not fulfilled sufficiently, the ammeter will not give the same indications for the same current if that current has been reached (a) by increasing from a smaller current, or (b) by decreasing from a larger current. In this case there is said to be hysteresis in the readings. Although therefore most simple and cheap to construct, such soft-iron instruments are not well adapted for accurate work. A much better form of electromagnetic ammeter can be constructed on a principle now extensively employed, which consists in pivoting in the strong field of a permanent magnet a small coil through which a part of the current to be measured is sent. Such an instrument is called a shunted movable coil ammeter, and is represented by a type of instrument shown in fig. 3. The

FIG. 3.--Shunted Movable Coil Ammeter, Isenthal & Co.

construction of this instrument is as follows:--Within the instrument is a horseshoe magnet having soft-iron pole pieces so arranged as to produce a uniform magnetic field. In this magnetic field is pivoted a small circular or rectangular coil carried in jewelled bearings, the current being passed into and out of the movable coil by fine flexible conductors. The coil carries an index needle moving over a scale, and there is generally an iron core in the interior of the coil but fixed and independent of it. The coil is so situated that, in its zero position when no current is passing through it, the plane of the coil is parallel to the direction of the lines of force of the field. When a current is passed through the coil it rotates in the field and displaces the index over the scale against the control of a spiral spring like the hairspring of a watch. Such instruments can be made to have equidivisional scales and to read from zero upwards. It is essential that the permanent magnet should be subjected to a process of ageing so that its field may not be liable to change subsequently with time.

In the case of ammeters intended for very small currents, the whole current can be sent through the coil, but for larger currents it is necessary to provide in the instrument a shunt which carries the main current, the movable coil being connected to the ends of this shunt so that it takes a definite small fraction of the current passed through the instrument. Instruments of this type with a permanent magnetic field are only available for the measurement of continuous currents, but soft-iron instruments of the above-described gravity type can be employed with certain restrictions for the measurement of alternating currents. Direct reading equidivisional movable coil ammeters can be made in various portable forms, and are very much employed as laboratory instruments and also as ammeters for the measurement of large electric currents in electric generating stations. In this last case the shunt need not be contained in the instrument itself but may be at a considerable distance, wires being brought from the shunt which carries the main current to the movable coil ammeter itself, which performs the function simply of an indicator,

3. Electrodynamic Ammeters.--Instruments of the third class depend for their action on the fact discovered by Ampere, that mechanical forces exist between conductors carrying electric currents when those conductors occupy certain relative positions. If there be two parallel wires through which currents are passing, then these wires are drawn together if the currents are in the same direction and pressed apart if they are in opposite directions. (See ELECTROKINETICS.) Instruments of this type are called Electrodynamometers, and have been employed both as laboratory research instruments and for technical purposes. In one well-known form, called a Siemens Electrodynamometer, there is a fixed coil (fig. 4), which is surrounded by another coil having its axis at right angles to that of the fixed coil. This second coil is suspended by a number of silk fibres, and to the coil is also attached a spiral spring the other end of which is fastened to a torsion head. If then the torsion head is twisted, the suspended coil experiences a torque and is displaced through

FIG. 4.--Siemens Electrodynamometer. F, Fixed coil; D, Movable coil; S, Spiral spring; T, Torsion head; MM, Mercury cups; I, Index needle.

an angle equal to that of the torsion head. The current can be passed into and out of the movable coil by permitting the ends of the coil to dip into two mercury cups. If a current is passed through the fixed coil and movable coil in series with one another, the movable coil tends to displace itself so as to bring the axes of the coils, which are normally at right angles, more into the same direction. This tendency can be resisted by giving a twist to the torsion head and so applying to the movable coil through the spring a restoring torque, which opposes the torque due to the dynamic action of the currents. If then the torsion head is provided with an index needle, and also if the movable coil is provided with an indicating point, it is possible to measure the torsional angle through which the head must be twisted to bring the movable coil back to its zero position. In these circumstances the torsional angle becomes a measure of the torque and therefore of the product of the strengths of the currents in the two coils, that is to say, of the square of the strength of the current passing through the two coils if they are joined up in series. The instrument can therefore be graduated by passing through it known and measured continuous currents, and it then becomes available for use with either continuous or alternating currents. The instrument can be provided with a curve or table showing the current corresponding to each angular displacement of the torsion head. It has the disadvantage of not being direct reading when made in the usual form, but can easily be converted into a direct reading instrument by appropriately dividing the scale over which the index of the torsion head moves.

Ampere Balance.--Very convenient and accurate instruments based on the above principles have been devised by Lord Kelvin, and a large variety of these ampere balances, as they are called, suitable for measuring currents from a fraction of an ampere up to many thousands of amperes, have been constructed by that illustrious inventor. The difficulty which has generally presented itself to those who have tried to design instruments on the

FIG. 5.--Kelvin Flexible Metallic Ligament.

electrodynometer principle for use with large currents has been that of getting the current into and out of the movable conductor, and yet permitting that conductor to remain free to move under very small force. The use of mercury cups is open to many objections on account of the fact that the mercury becomes oxidized, and such instruments are not very convenient for transportation. The great novelty in the ampere balances of Lord Kelvin was a joint or electric coupling, which is at once exceedingly flexible and yet capable of being constructed to carry with safety any desired current. This he achieved by the introduction of a device which is called a metallic ligament. The general principle of its construction is as follows:--Let +A, -A (fig. 5), be a pair of semi-cylindrical fixed trunnions which are carried on a supporting frame and held with flat sides downwards. Let +B, -B, be two smaller trunnions which project out from the sides of the two strips connecting together a pair of rings CC. The rings and the connecting strips constitute the circuit which is to be rendered movable. A current entering by the trunnion + B flows round the two halves of the circuit, as shown by the arrows, and comes out at the trunnion -B. In fig. 5 the current is shown dividing round the two rings; but in all the balances, except those intended for the largest currents, the current really circulates first round one ring and then round the other. To make the ligament, a very large number of exceedingly fine copper wires laid close together are soldered to the upper surface of the upper trunnion. The movable circuit CC thus hangs by two ligaments which are formed of very fine copper wires. This mode of suspension enables the conductor CC to vibrate freely like a balance, but at the same time very large currents can easily be passed through this perfectly flexible joint. Above and below these movable coils, which form as it were the two scale- pans of a balance, are fixed other stationary coils, and the connexions of all these six coils (shown in fig. 6) are such that when a current

FIG. 6.--Connexions of Kelvin Ampere Balance.

is passed through the whole of the coils in series, forces of attraction and repulsion are brought into existence which tend to force one movable coil upwards and the other movable coil downwards. This tendency is resisted by the weight of a mass of metal, which can be caused to slide along a tray attached to the movable coils. The appearance of the complete instrument is shown by fig. 7. When a current is passed through the instrument it causes one end of the movable system to tilt downwards, and the other end upwards; the sliding weight is then moved along the tray by means of a silk cord until equilibrium is again established. The value of the current in amperes is then obtained approximately by observing the position of the weight on the scale, or it may be obtained more accurately in the following

FIG. 7.--Lord Kelvin's Ampere Balance.

manner:--The upper edge of the shelf on which the weights slide (see fig. 8) is graduated into equal divisions, and the weight is provided with a sharp tongue of metal in order that its position on the shelf may be accurately determined. Since the current passing through the balance when equilibrium is obtained with a given weight is proportional to the square root of the couple due to this weight, it follows that the current strength when equilibrium is obtained is proportional to the product of the square root of the weight used

FIG. 8.--Slider of Kelvin Ampere Balance.

and the square root of the displacement distance of this weight from its zero position. Each instrument is accompanied by a pair of weights and by a square root table, so that the product of the square root of the number corresponding to the position of the sliding weight and the ascertained constant for each weight, gives at once the value of the current in amperes. Each of these balances is made to cover a certain range of reading. Thus the centi-ampere balance ranges from 1 to 100 centi-amperes, the deci-ampere balance from 1 to 100 deci-amperes, the ampere balance from 1 to 100 amperes, the deka-ampere balance from 1 to 100 amperes, the hecto-ampere balance from 6 to 600 amperes, and the kilo-ampere balance from 100 to 2500 amperes. They are constructed for the measurement not only of continuous or unvarying but also of alternating currents. In those intended for alternating currents, the main current through the movable coil, whether consisting of one turn or more than one turn, is carried by a wire rope, of which each component strand is insulated by silk covering, to prevent the inductive action from altering the distribution of the current across the transverse section of the conductor. To avoid the creation of induced currents, the coil frames and the base boards are constructed of slate. Kelvin ampere balances are made in two types--(1) a variable weight type suitable for obtaining the ampere value of any current within their range; and (2) a fixed weight type intended to indicate when a current which can be varied at pleasure has a certain fixed value. An instrument of the latter type of considerable accuracy was designed by Lord Kelvin for the British Board of Trade Electrical Laboratory, and it is there used as the principal standard ampere balance. A fixed weight is placed on one coil and the current is varied gradually until the balance is just in equilibrium. In these circumstances the current is known to have a fixed value in amperes determined by the weight attached to the instrument.

Calibration.--The calibration of ammeters is best conducted by means of a series of standard low resistances and of a potentiometer (q.v..) The ammeter to be calibrated is placed in series with a suitable low resistance which may be .1 ohm, .01 ohm, .001 ohm or more as the case may be. A steady continuous current is then passed through the ammeter and low resistance, placed in series with one another and adjusted so as to give any required scale reading on the ammeter. The potential difference of the ends of the low resistance is at the same time measured on the potentiometer, and the quotient of this potential difference by the known value of the low resistance gives the true value of the current passing through the ammeter. This can be then compared with the observed scale reading and the error of the ammeter noted.2

A good ammeter should comply with the following qualifications:--(1) its readings should be the same for the same current whether reached by increasing from a lower current or decreasing from a higher current; (2) if used for alternating currents its indications should not vary with the frequency within the range of frequency for which it is likely to be used; (3) it should not be disturbed by external magnetic fields; (4) the scale divisions should, if possible, be equal in length and there should be no dead part in the scale. In the use of ammeters in which the control is the gravity of a weight, such as the Kelvin ampere balances and other instruments, it should be noted that the scale reading or indication of the instrument will vary with the latitude and with the height of the instrument above the mean sea-level. Since the difference between the acceleration of gravity at the pole and at the equator is about 1/2%, the correction for latitude will be quite sensible in an instrument which might be used at various times in high and low latitudes. If G is the acceleration of gravity at the equator and g that at any latitude l, then g = G (1 + 0.00513 sin2 l). In the case of an instrument with gravity control, the latitude at which it is calibrated should therefore be stated.

FIG. 9.-- Edgewise Switchboard Ammeter, Kelvin & James White Ltd.

Switchboard Ammeters.--For switchboard use in electric supply stations where space is valuable, instruments of the type called edgewise ammeters are much employed. In these the indicating needle moves over a graduated cylindrically shaped scale, and they are for the most part electromagnetic instruments (see fig. 9).

BIBLIOGRAPHY.--Lord Kelvin (Sir W. Thomson), ``New Standard and Inspectional Electrical Measuring Instruments,'' Proc. Soc. Telegraph Engineers, 1888, 17, p. 540; J. A. Fleming, A Handbook for the Electrical Laboratory and Testing Room (2 vols., London, 1901, 1903 ); G. D. Aspinall Parr, Electrical Measuring Instruments (Glasgow, 1903); J. Swinburne, ``Electric Light Measuring instruments,'' Proc. Inst. Civ. Eng., 1891-1892, 110, pt. 4; K. Edgcumbe and F. Punga, ``Direct Reading Measuring Instruments for Switchboard Use,'' Jour. Inst. Elec. Eng., 1904, 33, p. 620. (J. A. F.)

1 See J. A. Fleming, A Handbook for the Electrical Laboratory and Testing Room, vol. i. p. 341 (1901), also A. Gray, Absolute Measurements in Electricity and Magnetism, vol. ii. pt. ii. p. 412 (1893).

2 See ``The Electrolysis of Copper Sulphate in Standardizing Electrical Instruments,'' by A. W. Meikle, read before the Physical Society of Glasgow University on the 27th of January 1888, or J. A. Fleming, A Handbook for the Electrical Laboratory and Testing Room, vol. i. p. 343.

AMPERSAND (a corruption of the mixed English and Latin phrase, ``and per se and,'' of which there are many dialect forms, as ``ampussyand,'' or ``amperseand''), the name of the sign & or &, which is a combination of the letters e, t, of the Lat. et= and. The sign is now usually called ``short and.'' In old-fashioned primers and nursery books the name and sign were always added at the end of the alphabet.

AMPHIARAUS, in Greek mythology, a celebrated seer and prince of Argos, son of Oicles (or Apollo) and Hypermestra, and through his father descended from the prophet Melampus (Odyssey, xv. 244). He took part in the voyage of the Argonauts and in the chase of the Calydonian boar; but his chief fame is in connexion with the expedition of the Seven against Thebes, organized by Adrastus, the brother of his wife Eriphyle, for the purpose of restoring Polyneices to the throne. Amphiaraus, foreseeing the disastrous issue of the war, at first refused to share in it; he had, however, promised Eriphyle when he married her that, in the event of any dispute arising between her brother and himself, she should decide between them; and now Eriphyle, bribed by Polyneices with the fatal necklace given by Cadmus to Harmonia, persuaded him against his better judgment to set out on the expedition. Knowing his doom, he bade his sons, Alcmaeon and Amphilochus, avenge his death upon their mother, upon whom, as he stepped into his chariot, he turned a look of anger. This scene was represented upon the chest of Cypselus described by Pausanias (v. 17).

The assault on Thebes was disastrous for the Seven; and Amphiaraus, pursued by Periclymenus, would have been slain with his spear, had not Zeus with a thunderbolt opened a chasm into which the seer, with his chariot, horses and charioteer, disappeared. Henceforth he was numbered with the immortals and worshipped as a god. Near Oropus, on the supposed site of his passing, his sanctuary arose, with healing springs, and an oracle famous for its interpretation of dreams (Pausanias i. 34). The ruins of this temple, with inscriptions which identify it, have been discovered and preserved at Mavrodilisi, in the provinces of Boeotia and Attica. There was another temple dedicated to him on the road from Thebes to Potniae, and here was the oracle of Amphiaraus consulted by Croesus and Mardonius.

Homer, Odyssey, xi. 326; Herodotus viii. 134; Pindar, Olympia, vi., Nemea, ix.; Apollodorus iii. 6.

AMPHIBIA, a zoological term originally employed by Linnaeus to denote a class of the Animal Kingdom comprising crocodiles, lizards and salamanders, snakes and Caeciliae, tortoises and turtles and frogs; to which, in the later editions of the Systema Naturae he added some groups of fishes. In the Tableau Elementaire, published in 1795, Cuvier adopts Linnaeus's term in its earlier sense, but uses the French word ``Reptiles,'' already brought into use by Brisson, as the equivalent of Amphibia. In addition Cuvier accepts the Linnaean subdivisions of Amphibia-Reptilia for the tortoises, lizards (including crocodiles), salamanders and frogs; and Amphibia-Serpentes for the snakes, apodal lizards and Caeciliae.

In 17991 Alexandre Brongniart pointed out the wide differences which separate the frogs and salamanders (which he terms Batrachia) from the other reptiles; and in 1804 P. A. Latreille,2 rightly estimating the value of these differences, though he was not an original worker in the field of vertebrate zoology, proposed to separate Brongniart's Batrachia from the class of Reptilia proper, as a group of equal value, for which he retained the Linnaean name of Amphibia.

Cuvier went no further than Brongniart, and, in the Regne Animal, he dropped the term Amphibia, and substituted Reptilia for it. J. F. Meckel,3 on the other hand, while equally accepting Brongniart's classification, retained the term Amphibia in its earlier Linnaean sense; and his example has been generally followed by German writers, as, for instance, by H. Stannius, in that remarkable monument of accurate and extensive research, the Handbuch der Zootomie (2nd ed., 1856).

In 1816, de Blainville,4 adopting Latreille's view, divided the Linnaean Amphibia into Squamiferes and Nudipelliferes, or Amphibiens; though he offered an alternative arrangement, in which the class Reptiles is preserved and divided into two subclasses, the Ornithoides and the Ichthyoides. The latter are Brongniart's Batrachia, plus the Caeciliae, whose true affinities had, in the meanwhile, been shown by A. M. C. Dumeril; and, in this arrangement, the name Amphibiens is restricted to Proteus and Siren.

B. Merrem's Pholidota and Batrachia (1820), F. S. Leuckart's Monopnoa and Dipnoa (1821), J. Muller's Squamata and Nuda (1832), are merely new names for de Blainville's Ornithoides and Ichthyoides, though Muller gave far better anatomical characters of the two groups than had previously been put forward.

Moreover, following the indications already given by K. E. von Baer in 1828,5 Muller calls the attention of naturalists to the important fact, that while all the Squamata possess an amnion and an allantois, these structures are absent in the embryos of all the Nuda. An appeal made by Muller for observations on the development of the Caeciliae, and of those Amphibia which retain gills or gill-clefts throughout life, has unfortunately yielded no fruits.

In 1825 P. A. Latreille6 published a new classification of the Vertebrata, which are primarily divided into Haematherma. containing the three classes of Mammifera, Monotremata and Aves; and Haemacryma, also containing three classes-- Reptilia, Amphibia and Pisces. This division of the Vertebrata into hot and cold blooded is a curiously retrograde step, only intelligible when we reflect that the excellent entomologist had no real comprehension of vertebrate morphology; but he makes some atonement for the blunder by steadily upholding the class distinctness of the Amphibia. In this he was followed by Dr J. E. Gray; but Dumeril and Bibron in their great work,7 and Dr Gunther in his Catalogue, in substance, adopted Brongniart's arrangement, the Batrachia being simply one of the four orders of the class Reptilia. Huxley adopted Latreille's view of the distinctness of the Amphibia, as a class of the Vertebrata, co-ordinate with the Mammalia, Aves, Reptilia and Pisces; and the same arrangement was accepted by Gegenbaur and Haeckel. In the Hunterian lectures delivered at the Royal College of Surgeons in 1863, Huxley divided the Vertebrata into Mammals, Sauroids and Ichthyoids, the latter division containing the Amphibia and Pisces. Subsequently he proposed the names of Sauropsida and Ichthyopsida for the Sauroids and Ichthyoids respectively.

Sir Richard Owen, in his work on The Anatomy of Vertebrates, followed Latreille in dividing the Vertebrata into Haematotherma and Haematocrya, and adopted Leuckart's term of Dipnoa for the Amphibia. T. H. Huxley, in the ninth edition of this Encyclopaedia, treated of Brongniart's Batrachia, under the designation Amphibia, but this use of the word has not been generally accepted. (See BATRACHIA.) (T. H. H.; P. C. M.)

1 Brongniart's Essai d'une classification naturelle des reptiles was not published in full till 1803. It appears in the volume of the Memoires presentes a l'Institut par divers savans for 1805.

2 Nouveau dictionnaire d'histoire naturelle, xxiv., cited in Latreille's Fannilles naturelles du regne animal.''

3 System der vergleichenden Anatomie (1821).

4 ``Prodrome d'une Nouvelle Distribution du regne Animal.'' Bulletin des sciences par la Societe Philomatique de Paris (1816), p. 113.

5 Entwickelungs-Geschichte der Thiere, p. 262

6 Familles naturelles du regne animal.

7 Erpetologie generale, ou histoire naturelle complete des reptiles (1836).

AMPHIBOLE, an important group of rock-forming minerals, very similar in chemical composition and general characters to the pyroxenes, and like them falling into three series according to the system of crystallization. They differ from the pyroxenes, however, in having an angle between the prismatic cleavage of 56 deg. instead of 87 deg. ; they are specifically lighter than the corresponding pyroxenes; and, in their optical characters, they are distinguished by their stronger pleochroism and by the wider angle of extinction on the plane of symmetry.

They are minerals of either original or secondary origin; in the former case occurring as constituents (hornblende) of igneous rocks, such as granite, diorite, andesite, &c. Those of secondary origin have either been developed (tremolite) in limestones by contact-metamorphism, or have resulted (actinolite) by the alteration of augite by dynamo-metamorphism. Pseudomorphs of amphibole after pyroxene are known as uralite.

The name amphibole (from the Gr. amfibolos, ambiguous) was used by R. J. Hauy to include tremolite, actinolite and hornblende; this term has since been applied to the whole group. Numerous sub-species and varieties are distinguished, the more important of which are tabulated below in three series. The formulae of each will be seen to conform to the general metasilicate formula R''SiO3.

ORTHORHOMBIC SERIES. Anthophyllite . . (Mg,Fe)SiO3. MONOCLINIC SERIES. Tremolite . . CaMg3(SiO3)4. Actinolire . . Ca(Mg,Fe)3(SiO3)4. Cummingtonite . (Fe,Mg)SiO3. Richterite . . (K2,Na2,Mg,Ca,Mn)SiO3. Hornblende . . {Ca(Mg,Fe)3(SiO3)4 with {NaAl(SiO3)2 and (Mg,Fe) (Al,Fe)2SiO6. MONOCLINIC SERIES--continued.

Glaucophane . . NaAl(SiO3)2.(Fe,Mg)SiO3. Crocidolite . . NaFe(SiO3)2.FeSiO3. Riebeckite . . 2NaFe(SiO3)2.FeSiO3. Arfvedsonite . . Na8(Ca,Mg)3(Fe,Mn)14(Al,Fe)2 Si21O45. ANORTHIC SERIES. Aenigmatite . . Na4Fe''9Al Fe'' '(Si,Ti)12O38.

Of these, tremolite, hornblende and crocidolite, as well as the important varieties, asbestos and jade, are treated under their own headings. Brief mention only need be here made of some of the others. Naturally, on account of the wide variations in chemical composition, the different members vary considerably in characters and general appearance; the specific gravity, for example, varies from 2.9 in tremolite to 3.8 in aenigmatite.

Anthophyllite occurs as brownish, fibrous or lamellar masses with hornblende in mica-schist at Kongsberg in Norway and some other localities. An aluminous variety is known as gedrite, and a deep green, Russian variety containing little iron as kupfferite.

Actinolite is an important member of the monoclinic series, forming radiating groups of acicular crystals of a bright green or greyish-green colour. It occurs frequently as a constituent of crystalline schists. The name (from aktis, a ray, and lithos, a stone) is a translation of the old German word Strahlstein, radiated stone.

Glaucophane, crocidolite, riebeckite and arfvedsonite form a somewhat special group of alkali-amphiboles. The two former are blue fibrous minerals occurring in crystalline schists, and are the result of dynamo-metamorphic processes; the two latter are dark green minerals which occur as original constituents of igneous rocks rich in soda, such as nepheline-syenite and phonolite.

Aenigmatite and its variety cossyrite are rare minerals forming constituents of igneous rocks of the nepheline-syenite and phonolite groups. (L. J. S.)

AMPHIBOLITE, the name given to a rock consisting mainly of amphibole (hornblende), the use of the term being restricted, however, to metamorphic rocks. Holocrystalline plutonic igneous rocks composed essentially of hornblende are known as hornblendites. As is the case with most petrological terms the exact connotation is not very strictly defined; most authors allow that accessory minerals such as felspar, garnet, augite and quartz may be present in variable and often considerable amount. A foliated or schistose structure, though often developed in these rocks, is not universal. The hornblende is usually dark green (actinolite) but may be nearly black in the hand specimen; in the microscopic slide it is commonly green of various shades, but may be brown, blue or nearly colourless. It frequently occurs in elongated bladed prisms, but rarely shows good crystal faces. The term hornblende-schist is employed by many writers as nearly synonymous with amphibolite; most hornblende-schists contain felspar and iron oxides, while sphene, rutile, quartz and apatite are rarely absent. Reddish garnets are often conspicuous in the rocks of this group (garnet-amphibolites), and when in addition a green-coloured augite occurs the rocks are intimately allied to the hornblende-eclogites. Epidote also, in yellow grains, is common (epidote-amphibolites), and in these rocks the hornblende may be of the blue and richly pleochroic variety known as glaucophane (glaucophane-epidote-schists). Hornblende-schists containing dark green ferriferous hornblende (grunorite-schists) are abundant in some parts of North America. Tremolite-schists consist essentially of white or very pale green amphibole; occasionally they are black from the presence of numerous minute grains of iron oxide or of graphite. Many tremolite-schists contain much talc and chlorite, and as these rocks have been derived from peridotites they not infrequently show residual grains of olivine. Nephrite (Gr. nefros, a kidney) is a very compact, hardly schistose amphibolite, consisting of fine interwoven fibres of hornblende. Among other accessory minerals biotite, chlorite, talc, scapolite and tourmaline may be mentioned; if abundant they give rise to special varieties such as biotite-amphibolite, &c.

The amphibolites are typical rocks of the metamorphic group and as such attain a large development in all regions of crystalline schists and gneisses such as the Alps, Ardennes, Harz, Scottish Highlands, and the Lakes district of North America. They occur in two ways, viz. as large circular or elliptical areas which mark the site of old plutonic stocks or bosses of basic rock, and as long narrow strips intercalated among outcrops of other metamorphic rocks. Regarded from the point of view of their origin they fall into two groups, the ortho-amphibolites, which are modified igneous rocks, and the para-amphibolites, which are altered sediments. The former are far the more common. Igneous rocks which contain much augite (e.g. dolerites, gabbros, diabases, pyroxenites and many peridotites) are usually converted into amphibolites when they are subjected to pressure and interstitial movements during earth-folding. If felspar be present also, epidote may form, while part of the felspar recrystallizes as a species of the same mineral richer in alkalies or as mica. Olivine and ilmenite, the other common constituents of these rocks, may, alone or in conjunction with the above-named minerals, yield garnet, talc, sphene, rutile, &c. There is little or no alteration in the bulk composition of the rock, but its component elements enter into new combinations. Chemical analysis, accordingly, will often enable us to identify an igneous rock (diabase, &c.) under the guise of an amphibolite. The transformation of the rock may be complete, so that no trace is left of the original structures or minerals. Very often, however, it is only partial, and by obtaining a sufficiently large number of specimens a series of intermediate or transitional stages may be studied; these prove conclusively the nature of the process, though its causes are less clearly understood. Green hornblende may be seen gradually replacing augite, at first in needle-like crystals, for which gradually more compact masses are substituted. The felspar breaks up into a mosaic in which albite, epidote or zoisite, quartz and garnet may often be identified. Biotite and primary hornblende suffer comparatively little change; olivine disappears, and garnet, talc and tremolite or anthophyllite take its place. The original structures of this group of rocks (ophitic, porphyritic, poikilitic, vesicular, &c.) gradually fade away, and merge into those of the metamorphic amphibolites. Even when the greater part of the rock mass has suffered complete reconstruction, kernels or phacoids may remain, showing the old igneous structures, though the minerals are greatly altered. The transitional stages from gabbro or diabase to amphibolite are so common that they form a widespread and important group of rocks, which have been described under the names greenstone, greenstone-schist, flaser-gabbro, saussurite- gabbro, meta-diabase, &c. The ortho-amphibolites also include a small group of igneous rocks, which have a foliated or banded structure due to movements and pressure during consolidation, e.g. foliated diorite or diorite-schist.

The sedimentary amphibolites or para-amphibolites, less common than those above described, are frequent in some districts, such as the northern Alps, southern highlands of Scotland, Green Mountains, U.S.A. Many of them have been ash-beds, and their conversion into hornblende-schists follows exactly similar stages to those exemplified by basic crystalline igneous rocks. Others have been greywackes of varied composition with epidote, chlorite, felspar, quartz, iron oxides, &c., and may have been mixed with volcanic materials, or may be partly derived from the disintegration of basic rocks. When they are most metamorphosed they are often very hard to distinguish from igneous hornblende-schists; yet they rarely fail to reveal signs of bedding, pebbly structure, sedimentary banding and gradual transition into undoubtedly sedimentary types of gneiss and schist. Deposits containing dolomite and siderite also readily yield amphibolites (tremolite-schists, grunorite-schists, &c.) especially where there has been a certain amount of contact metamorphism by adjacent granitic masses. (J. S. F.)

AMPHIBOLOGY, or AMPHIBOLY (Gr. ampibolia), in logic, a verbal fallacy arising from ambiguity in the grammatical structure of a sentence (Aristot., Organon,Soph., El., chap. iv.). It occurs frequently in poetry, owing to the alteration for metrical reasons of the natural order of words; Jevons quotes as an example Shakespeare, Henry VI.: ``The duke yet lives that Henry shall depose.''

AMPHICTYONY (Gr. amfiktuonia, i.e. a body composed of amfiktiones, amfiktuones, ``dwellers around''), an association of ancient Greek communities centring in a shrine. As the extant sources do not define the term, and as they apply it to but five or six associations, the majority of which are little known, modern scholars are in doubt as to the essential character of the institution, and hesitate therefore to extend the name beyond this limited list. The word itself indicates that the association primarily comprised neighbours, though the Delphic amphictyony came in time to include relatively distant communities (Strabo ix. 3, 7). For the origin of the institution it is safe to assume that neighbouring communities, whether tribes (ethne) or cities, desiring friendly intercourse with one another chose the sanctuary of some deity conveniently situated, at which to hold their periodical festival for worship and their fair for the interchange of goods. If the limited use of the word according to our sources is not purely accidental, at all events there were many Greek leagues, not expressly termed amphictyonies, which had the characteristics here stated.

The Delian amphictyony probably reached the height of its splendour early in the 7th century B.C. The Hymn to the Delian Apollo, composed about that time, celebrates the gathering of the Ionians with their wives and children at the shrine of their god on the island of Delos, to worship him with music, dancing and gymnastic contests (vv. 146-164; cf. Thuc. iii. 104). The later misfortunes of the Ionians caused a decline of the festival. Peisistratus, taking possession of Delos, seems to have used the sanctuary as a means of extending his political influence. When after the great war with Persia the Aegean cities under the leadership of Athens united in a political league (477 B.C.), they chose as its centre the temple of the Delian Apollo, doubtless through a desire to connect the new alliance with the associations of the old amphictyony. How far the council and other institutions of the Delian confederacy were based upon the amphictyonic organization cannot be determined. The removal of the treasury to Athens in 454 B.C. deprived Delos of political importance, though the amphictyony continued. The council gradually dwindled, and probably came to an end without formal abolition. In 426 B.C. the Athenians purified the island and instituted a great festival to be held under their presidency every four years (Thuc. iii. 104). In 422 they expelled the Delians (Thuc. v. 1). At the end of the Peloponnesian War Athens was deprived of Delos along with her other possessions, but she appears to have regained control of the island after the victory of Cnidus (394). An inscription of 390 B.C. proves that at this date Athenian authority had been restored. The affairs of the temple were managed by a board of five Athenian amphictyons, assisted by some Delian officials (inscrr. in Bull. Hell. viii. 284, 304, 307 f.); and in the 4th century we again hear of a council in addition to the board (CIG. i. 158). At this time the amphictyony is known to have embraced both the Athenians and the inhabitants of the Cyclades; but a strong Delian party bitterly opposed Athenian rule (cf. inscr. in Bull. Hell. iii. 473 f.), which came to an end with the supremacy of Macedon. The dissolution of the amphictyony soon followed.

Far more famous is the Delphic, or more strictly, the Pylaeic-Delphic, amphictyony. It was originally composed of twelve tribes dwelling round Thermopylae--the Thessalians, Boeotians, Dorians, Ionians, Perrhaebians, Magnetes, Locrians, Oetaeans, Phthiotes, Mahans, Phocians (Aeschin. ii. 116), and Dolopians (Paus. x. 8. 2). The name of the council (pylaea) and of one set of deputies (pylagori), together with the important place held in the amphictyony by the temple of Demeter at Anthela, near Thermopylae, suggests that this shrine was the original centre of the association. How and when Delphi became a second centre is quite uncertain. The council of the league included deputies of two different kinds--pylagori and hieromnemones. the latter were twenty-four in number, two from each tribe. As the league was originally made up of neighbours, the Dorian tribe must have comprised simply the inhabitants of Doris; the Locrians were probably the eastern (Opuntian) branch; and the Ionians were doubtless limited to the adjacent island of Euboea. Afterwards, by affiliating themselves to Doris, the Peloponnesian Dorians gained admission, and Athens must have entered as an Ionian city before the first Sacred War. Henceforth Athens monopolized one of the two Ionian votes, while the other passed in rotation among the remaining Ionic, perhaps only among the Euboeic, cities. In the same way Doris held one Dorian vote and the other passed in rotation among the Dorian cities of Peloponnesus; and the east and west Locrians came to have one each. When after the second Sacred War the Phocians were expelled, Macedon received their two votes (346 B.C.) About the same time the Perrhaebians and the Dolopians were deprived of half their representation, and the two votes were transferred to the Delphians (inscrr. in N. Jahrb. f. cl. Philol. clv. 742, cf. 743, 753; Bull. Hell. xxi. 322, cf. 325; Bourguet, Sanct. Pyth. 145, 147). In the following century the Aetolians gained such dominance in the amphictyony as to convert the council into an organ of their league. Recent research has made it appear certain (cf. Pomptow, ib. 754 ff.) that they were never formally admitted to membership, but that they maintained their supremacy in the council (Livy xxxi. 32. 3; Polyb. iv. 25. 8) by controlling the votes of their allies, who-- called Aetolians in the inscriptions--were often in the majority. They made no material change in its composition, which, accordingly, after the dissolution of their league by the Romans is found to be nearly as it was after the second Sacred War. A few minor changes came in under the supremacy of the Roman republic; and finally Augustus increased the number of votes to thirty, and distributed them according to his pleasure. In the age of the Antonines the association was still in existence (paus. x. 8. 4 f.).

Although the hieromnemones of the Thessalians, who held the presidency, and perhaps of a few other communities, must have been elected, the office was ordinarily, as at Athens, filled by lot. As a rule they were renewed annually (Aristoph. Clouds, 623 f.; Foucart, in Bull. Hell. vii. 411, 413 f.). Each hieromnemon was accompanied by two pylagori, elected semi-annually (Demosth. xviii. 149; Aeschin. iii. 115; Tim. Lex. Plat., s.v. 'Amfiktuones), and representing the same tribe, though not necessarily the same city. On one occasion Athens is known to have sent three. The hieromnemones were formally superior, but because of the method of appointment they were necessarily men of mediocre ability, inexperienced in speaking and public business, and for that reason they readily became the tools of the pylagori, who were orators and statesmen. In the literary sources, accordingly, the latter are rightly given credit for the acts of the council; it was the pylagori who set a price on the head of the traitor Ephialtes ( Herod. vii. 213 ), and who on the motion of Themistocles rejected the proposition of Lacedaemon for the expulsion of the states which had sided with Persia (Plut. Them. 20). The pylagori had a right to propose measures and to take part in the deliberations; they as well as the hieromnemones were required to take the juror's oath; and the acts of the council were inscribed officially as resolutions of the hieromnemones and pylagori conjointly. The hieromnemon, however, cast the vote of his community, though in the record his two pylagori were made equally responsible for it. The necessary inference from these facts is that the vote was determined by a majority of the three deputies (inscr. in Bull. Hell. xxvii. 106-111, A 20-33; B 1-10). The council decided all questions which fell within its competence. Matters of greater importance, as the levy of an extraordinary fine on a state or the declaration of a sacred war, it presented in the form of a resolution to an assembly (ekklesia), composed of the deputies, the amphictyonic priests, and any other citizens of the league who chanced to be present (Aeschin. iii. 124; cf. Hyp. iv. 7, 26 f.). This assembly was relatively unimportant, however, and is mentioned only by the two authorities here cited.

It is now well established by epigraphic evidence (Bull. Hell. vii. 412 f., 417; Pomptow, in N. Jahrb. f. cl. Philol. cxlix. 826-829) that the amphictyons met both in the spring and in the autumn at Delphi, and the literary sources should alone be sufficient authority for meetings in the same seasons at Thermopylae (Hyp. iv. 7, 25 ff.; Strabo ix. 3, 7, 4, 17; Harpocration, s.v. Pulai.) It is known, too, that the meeting at Thermopylae followed that at Delphi (inscr. in Bull. Hell. xxiv. 136 f.).

The primary function of the council was to administer the temporal affairs of the two shrines, of which the sanctuary of Apollo at Delphi claimed by far the greater share of attention. The hieromnemones were required periodically to inspect the lands belonging to this god, to punish those who encroached, and to see that the tenants rendered their quota of produce; and the council held the states responsible for the right performance of such duties by their respective deputies (CIA. ii. 545; inscr. in Bull. Hell. vii. 428 f.). Another task of the council was to supervise the treasury, to protect it from thieves, and by investments to increase the capital (Strabo ix. 3, 7; Isoc. xv. 232; Demosth. xxi. 144; Plut. Sull. 12). Naturally, too, it controlled the expenditure. We find it, accordingly, in the 6th century B.C. contracting for the rebuilding of the Delphic temple after it had been destroyed by fire (Herod. v. 62; Paus. x. 5. 13), and in the 4th century creating an Hellenic college of temple-builders for the purpose (inscrr. in Bull. Hell. xx. 202 f., 206, xxi. 478, xxiv. 464), adorning the interior with statues and pictures (Diod. xvi. 33), inscribing the proverbs of the Seven Sages on the walls (Paus. x. 24. 1), bestowing crowns on benefactors of the god (CIG. i. 1689 b), preparing for the Pythian games, awarding the prizes (Pind. Pyth. iv. 66, x. 8 f.), instituting a board of treasurers (inscr. in Bourguet, Sanct. Pyth. 175 ff.) and issuing coins. It was also in the material interest of Apollo that the council passed a law which forbade the Greeks to levy tolls on pilgrims to the shrine (Aeschin. iii. 107; Strabo ix. 3, 4), and another requiring the amphictyonic states to keep in repair their own roads which led towards Delphi (CIA. ii. 545). A law of great interest, dating from the beginning of the institution, imposed an oath upon the members of the league not to destroy an amphictyonic city or to cut it off from running water in war or peace; but to wage war upon those who transgressed this ordinance, to destroy their cities, and to punish any others who by theft or plotting sought to injure the god (Aeschin. ii. 115). In this regulation, which was intended to mitigate the usages of war amongst the members of the league, we have one of the origins of Greek interstate law. Though other regulations were made to secure peace at the time of the festival (Dion. Hal. iv. 25. 3), and though occasionally the council was called upon to arbitrate in a dispute (cf. Demosth. xviii. 135), no provision was made to compel arbitration.