Page images

atom we do not find this to be the case, while with Rb and Cs the Diffuse series has not been observed yet.

by 30 give the numbers 59


In the Principal series of the Alkalis the values of μ when written 1.965 2.1 2.233 2-3 2-3 and then multiplied 67 69 69 the differences between which are 4 4 2 0. In other words, the value of for Li can be written 2 1/30, and those for the other metals of the family be obtained by adding in succession 4 4 2 and 0 thirtieths. Moreover the values of μ in the Sharp series can be derived from those in the Principal series by sub


[blocks in formation]


[ocr errors]


12 +30


Concerning for the other metals, it may be worth noting that in many cases the values for the Diffuse and Sharp series when added together give a sum near 1·08, the actual values







Mg. Hg. 1.06 1.098 1.108 1.044 1.101 1.077 1.098

In the special Magnesium series of section 2, where μ has. the form -2.343/m, which may possibly be 2·333/m or 2÷m, we have evidence of stationary waves in the atom caused by 2 times the circumference being divided into 3, 4,..... 8 standing waves, each of which is capable of maintaining a corresponding motion of an electron.

For the value of 1/A we have VB with V=3× 1010 and B for hydrogen 109675, giving 1/A=33 × 1014 revolutions per second. Now this is of the same order as 102 times the frequency of the mechanical vibrations of the Li atom and as 10 times the frequency of ordinary light. It is therefore probable that the frequency of the revolution of the electrons in a neutron is the same as that frequency which is common to the mechanical vibrations of all the atoms, for μ in the Li family bears witness to harmonics of frequency 30 times the fundamental. Atom and electron by their mutual resonance make this frequency of 33 × 101 the fundamental datum in the vibrations of atoms and electrons.

The reason why solids and liquids in general give continuous spectra is that in them the frequency of the collisions of an atom with others is of the order of the frequency of the internal vibrations of the atom, so that every mode of motion. has its energy renewed so often with sudden changes of phase, that there is no chance for the motions maintained by resonance to stand out from the crowd.

8. Summary.

The structure of spectra is traced to the two facts, first that atoms vibrate as deformable, but practically incompressible, bodies of finite calculable rigidity, so that their surfaces have stationary waves corresponding to the fundamental mode of vibration and its harmonics, and second that electrons in describing nearly circular orbits round an atom, out of an infinity of such orbits possible, have orbits of certain frequencies made predominant by resonance. An electron can make 1, 2, 3 .... m revolutions between two occasions when it gets its energy renewed by striking the atom at the middle of one of its vibrating internodes, or it can make 1+μ, 2 + μ.... m + u where u is generally an harmonic fraction. The orbits for a positive electron are different from those for a negative, and therefore a relative motion between positive and negative electron is set up. This is the direct cause of the vibrations of light. This relative motion can be represented by giving the positive and negative electron different angular velocities round the circumference of a circle. By such motions Balmer's and Rydberg's formulas can be explained, and Rydberg's laws lead to the conclusion that the fundamental angular velocity of all electrons associated with all atoms is a constant representing a frequency of 33 x 1014 per second.



From the rigidities of the metals at absolute zero the mechanical periods of vibration of the atoms are calculated and proved to exhibit simple harmonic relations, and probably to possess a common harmonic of a frequency of the same order as that of ordinary light. It is probable that the common harmonic of the atoms and the fundamental mode of motion of the electron are identical or harmonically related.

The spectra of different elements thus appear to be caused by practically one and the same form of electrical appliance (pair of electrons) which is supplied with energy by the atom at various internodes. One spectrum is only a slight kinematical variation of another.

The principles of the kinetic theory of gases and of the electromagnetic theory of light are brought into natural relation. Rigidities at absolute zero calculated according to the Kinetic Theory of Solids are shown to be connected with the structure of spectra, and molecular resonance is shown to play a striking part in the melting of the metals. The atoms in a compound molecule are more intimately united than is usually supposed.) Melbourne, May 1901.

XXIV. Notes on the Zeeman Effect.
By N. A. KENT*.

IT has been shown by H. M. Reesef that the separation of the external components of the regular Zeeman triplet or quadruplet, as seen perpendicular to the lines of force, does not vary proportionally with the strength of the magnetic field in which the luminous source is placed. This fact was established for various zinc and cadmium lines up to a field of about 26,000 c.G.S. units.

Reese also states, in referring to certain lines in the spectrum of iron, that," In comparing the separation of the lines between 3900 and 4450 it was at once observed that the lines could be broken up into two classes, in each of which the separation of the various lines was of the same magnitude. These two classes are identical with those for which Humphreys found that the shift due to pressure was the same. On these plates the separation is very small in all cases, owing to a weak field, and no accurate measurements were taken of the


It appeared then to be a matter of no little interest to extend Reese's investigations on zinc using higher fieldstrengths; and also to make a more exhaustive investigation of the spectrum of iron and measure the separation with care. These two primary lines of study suggested others as given


The apparatus used was essentially that employed by Reese-A Rowland concave grating, radius of curvature 13 feet 3 inches, of 15,000 lines per inch, fitted with the ordinary slit and camera box; Seed's" Gilt Edge," Cramer's "Isochromatic Fast," and the International Colour Photo. Co.'s "Erythro" plates; an electromagnet giving a maximum field of 33,000 c.G.s. units for a 3 mm. gap; as luminous Source, a spark between terminals of, or containing, the metals investigated-the spark being produced by an alternating current, of 133 cycles per second passed through an adjustable impedance and through a transformer, the secondary circuit containing a condenser which discharged across the spark-gap, the leads to which were short thick wires (self-induction was used in the discharge circuit when it was desirous to remove the air-spectrum or sharpen the lines of the metal under investigation); and a dividing engine, whose

From the Johns Hopkins University Circulars, vol. xx. no. 152 (May-June 1901).

† Astrophys. Journal, xii. No. 2, Sept. 1900, pp. 120–135.

screw was exceedingly accurate, used to measure the Zeeman separation.

Briefly the results of the investigation are as follows:

1. The separation of the outer components of the zinc lines 4680 38, 4722-26, 4810-71 is not proportional to the strength of field for values of the latter from 26,000 to 33,000 C.G.S. units; that is, Reese's results for zinc are confirmed and extended.

2. Further, for iron lines chosen somewhat at random the same is true, and the lines which, in the ordinary iron spectrum, are "nebulous" in character show the least increase in Zeeman separation as the field is increased in strength; or, in other words, the curve plotted between strength of field "H," and resulting separation "AN," shows the greatest droop.

3. Hence it follows that Becquerel and Deslandres *, who used a field of 35,000 c.G.s. units, were unjustified in their attempt to discover in the lines of the iron spectrum a law governing the separation. No such law as that proposed by them is apparent from measurements made on my plates of 80 lines which appeared in good form for measurement.

4. As to the pressure shift for iron :-In the spectrum of this metal 34 lines were investigated with great care. 26 show both large pressure shift and large separation, or small pressure shift and small separation; while & show either large pressure shift and small separation, or small pressure shift and large separation. Thus it cannot be said that, if a line show large pressure shift, it will show large separation; or, if small pressure shift, small separation.

5. Nickel and cobalt were investigated. No law governing the separation is apparent.

[merged small][merged small][ocr errors]

of Kayser and Runge's spectroscopic series was extended. In the expression given, "A" represents the Zeeman separation," the wave-length, and "H" the field strength. Preston deduced this law from measurements upon the lines in the 2nd subordinate series of cadmium and magnesium, whose wave-lengths are given by putting "n" equal to 3 in Kayser and Runge's formula :

[ocr errors]

where "X" is the wave-length, "A," "B," and "C" are constants, and "n" has integer values 3, 4, 5, &c. The lines

[blocks in formation]
[blocks in formation]

Zn. 3252-42 3303-03 3345-62 4680-38 4722-26 4810-71 Cd. 3403-74 3467-76 3613-04 4678.37 4800.00 5086-06 4046-78 4358-56 5460-97 3838-44 5167-55 5172-87 5783-84




[ocr errors]


6122-46 6162.46 4425 61 4435.86 4456-08

[merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][ocr errors][merged small][ocr errors][ocr errors][ocr errors][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small]

(1) Field 26,460 c.G.s. units.

(2) Reese's values calculated from slope of curve on H diagrams. (3) Approximate mean value given by Preston for the homologous lines of Zn, Cd, and Mg. H=20,000.

(4) Calculated from data given by Reese. See his article before men

tioned, Astrophys. Journ. No. 2, Sept. 1900, p. 129.

(5) Excluding strontium as it is irregular, and calcium, the measurements of which are but approximate. Note that the calcium lines agree quite well with the homologous lines in Zn, Cd, Hg, and Mg. Phil. Mag. S. 6. Vol. 2. No. 9. Sept. 1901.


« PreviousContinue »