Elements and Atoms: Chapter 11
An Unsystematic Foreshadowing: J. A. R. Newlands

John Alexander Reina Newlands (1837-1898; see photo at the Edgar Fahs Smith collection, University of Pennsylvania) was born in Southwark, England. He briefly attended the Royal College of Chemistry and was an assistant there. He spent most of his life in and around London, working as a private analytical chemist, as a teacher of chemistry, and for many years as the chief chemist at a sugar refinery.

Opinions concerning the amount of credit which Newlands deserves for uncovering the periodic law varied greatly during his lifetime and continue to do so. Newlands believed his own contributions were slighted. As will be seen below, his work was ridiculed when presented, and was not published by the Chemical Society. During the 1870s and 1880s while Mendeleev and Meyer were receiving accolades for their work Newlands asserted his own priority in print several times. One of these was in a letter to the Chemical News published within a month of the award of the Davy Medal to Mendeleev and Meyer [Newlands 1882] and the most extensive in a booklet which reprinted his contributions and several of his claims of priority [Newlands 1884]. A reviewer of this booklet opined that Newlands deserves credit for "the idea of periodicity among the elements," if not for the actual system expressing that periodicity [Chemical News 1884]. Yet by the end of Newlands' life an obituary would claim that "all who have taken the trouble to look into the literature of the subject [the periodic law], know that it was he who discovered the fundamental relation embodied in this so-called law" [W. A. T. 1898].

Twentieth-century judgments echo Newlands' contemporaries. J. W. van Spronsen's comprehensive study of the origins and development of the periodic law credits six researchers with independent discovery of the periodic system [van Spronsen 1969]; see also Taylor 1949. Yet chemists still ask, as in a recent article emphasizing Mendeleev's contributions, "is it any wonder that some of Newlands' contemporaries failed to take him seriously?" [Gorin 1996] I tend to agree with the contemporary review of Newlands' booklet, that Newlands deserves credit for the idea of a periodic system, but not for the system itself. He wrote about many of the features of the modern periodic classification, but often held those features in contradictory ways; as a result, his system was less than the sum of its parts. [Giunta 1999]

Having revealed where I stand, I invite you to examine and assess Newlands' work on the subject. This selection includes four short papers of Newlands and a report of another paper which show him struggling toward and eventually formulating the system he dubbed the "law of octaves."


On Relations among the Equivalents[1]

J. A. R. N., Chemical News 7, 70-72 (Feb. 7, 1863).

To the Editor of the CHEMICAL NEWS.

SIR,--Many chemists, and M. Dumas[2] in particular, have, on several occasions, pointed out some very interesting relations between the equivalents of bodies belonging to the same natural family or group[3]; and my present purpose is simply to endeavour to proceed a little further in the same direction. I must, however, premise that many of the observations here collected together are well known already, and are only embodied in my communication for the purpose of rendering it more complete.

Before proceeding any further, I may also remark, that in the difficult task of grouping the elementary bodies, I have been guided more by chemical characteristics than by physical appearances, and have, therefore, taken no notice of the ordinary distinction between metals and non-metallics. The numbers which I have attached to the various groups are merely for the purpose of reference, and have no further significance whatever. For the sake of perspicuity, I have employed the old equivalent numbers, these atomic weights being, with one or two exceptions, taken from the 8th edition of "Fownes' Manual."[4]

The following are among the most striking relations observed on comparing the equivalents of analogous elements. (In order to avoid the frequent repetition of the word "equivalent," I have generally used the names of the different elements as representing their equivalent numbers--thus, when I say that zinc is the mean of magnesium and cadmium, I intend to imply that the equivalent of zinc is the mean of those of magnesium and cadmium, and so on, throughout the paper):--

Group I. Metals of the alkalies:[5]-- Lithium, 7; sodium, 23; potassium, 39; rubidium, 85, caesium, 123; thallium, 204.

The relation among the equivalents of this group (see CHEMICAL NEWS, January 10, 1863) may, perhaps, be most simply stated as follows:--

1 of lithium + 1 of potassium = 2 of sodium.
1 " + 2 " = 1 of rubidium.
1 " + 3 " = 1 of caesium.
1 " + 4 " = 163, the equivalent of a metal not yet discovered.[6]
1 " + 5 " = 1 of thallium.
Group II. Metals of the alkaline earths:[7]-- Magnesium, 12; calcium, 20; strontium, 43.8; barium, 68.5.

In this group, strontium is the mean of calcium and barium.

Group III. Metals of the earths:[8]-- Beryllium, 6.9; aluminium, 13.7; zirconium, 33.6; cerium, 47; lanthanium, 47; didymium[9], 48; thorium, 59.6.

Aluminium equals two of beryllium, or one-third of the sum of beryllium and zirconium. (Aluminium also is one-half of manganese, which, with iron and chromium, forms sesquioxides, isomorphous, with alumina.)

1 of zirconium + 1 of aluminium = 1 of cerium.
1 " + 2 " = 1 of thorium.
Lanthanium and didymium are identical with cerium, or nearly so.

Group IV. Metals whose protoxides are isomorphous with magnesia:[10]-- Magnesium, 12; chromium, 26.7; manganese, 27.6; iron, 28; cobalt, 29.5; nickel, 29.5; copper, 31.7; zinc, 32.6; cadmium, 56.

Between magnesium and cadmium, the extremities of this group, zinc is the mean. Cobalt and nickel are identical. Between cobalt and zinc, copper is the mean. Iron is one-half of cadmium. Between iron and chromium, manganese is the mean.

Group V.[11]-- Fluorine, 19; chlorine, 35.5; bromine, 80; iodine, 127.

In this group, bromine is the mean between chlorine and iodine.

Group VI.[12]-- Oxygen, 8; sulphur, 16; selenium, 39.5; tellurium, 64.2.

In this group selenium is the mean between sulphur and tellurium.

Group VII.[13]-- Nitrogen, 14; phosphorus, 31; arsenic, 75; osmium, 99.6; antimony, 120.3; bismuth, 213.

In this group arsenic is the mean between phosphorus and antimony.

Osmium approaches the mean of arsenic and antimony, and is also almost exactly half the difference between nitrogen and bismuth, the two extremities of this group; thus, (213-14)/2 = 99.5.

Bismuth equals 1 of antimony + 3 of phosphorus; thus, 120.3 + 93 = 213.3.

Group VIII.[14]-- Carbon, 6; silicon, 14.20; titanium, 25; tin, 58.

In this group the difference between tin and titanium is nearly three times as great as that between titanium and silicon.

Group IX.[15]-- Molybdenum, 46; vanadium, 68.6; tungsten, 92; tantalium, 184.

In this group vanadium is the mean between molybdenum and tungsten.

Tungsten equals 2 of molybdenum, and tantalium equals 4 of molybdenum.

Group X.[16]-- Rhodium, 52.2; ruthenium, 52.2; palladium, 53.3; platinum, 98.7; iridium, 99.

In this group the first three are identical, or nearly so, and are rather more than half of the other two. (I may mention, by the way, that platinum is rather more than the half of gold; thus 98.7 x 2 = 197.4, gold being 197.)

Group XI.[17]-- Mercury, 100; lead, 103.7; silver, 108.

Lead is here the mean of the other two.

If we deduct the member of a group having the lowest equivalent from that immediately above it, we frequently observe that the numbers thus obtained bear a simple relation to each other, as in the following examples:--

Member of group havingOne immediately aboveDifference.
lowest equivalent.the preceding.
Magnesium 12Calcium 208
Oxygen 8Sulphur 168
Carbon 6Silicon 14.28.2
Lithium 7Sodium 2316
Fluorine 19Chlorine 35.516.5
Nitrogen 14Phosphorus 3117

A similar relation, though not quite so obvious as the above, may be shown by deducting the lowest member of a triad from the highest. The numbers thus obtained in the different triads correspond to a great extent. (By a triad I understand a group of analogous elements, the equivalent of one of which is the mean of the other two.) Of this relation I append a few examples:--

Lowest term of triad.Highest term of triad.Difference.
Lithium 7Potassium 3932
Magnesium 12Cadmium 5644
Molybdenum 46Tungsten 9246
Sulphur 16Tellurium 64.248.2
Calcium 20Barium 68.548.5
Phosphorus 31Antimony 120.389.3
Chlorine 35.5Iodine 12791.5

In the relation previously pointed out, the difference between the lowest member of a group, and the next above it, was either 8, or 8 x 2 = 16; and in the first of these triads the difference is 8 x 4 = 32; in the next four it approaches 8 + 6 = 48 [sic; obviously 8 x 6 is intended--CJG]; and in the last two triads it is nearly twice as great.[18]

The difference between the highest member of the platinum group, viz., iridium 99, and the lowest, rhodium 52.2, is 46.8, a number which approximates very closely to those obtained in some of the above triads; and it, therefore, appears possible that the platinum metals are the extremities of a triad, the central term or mean of which is at present unknown.[19]

I am, &c.

J. A. R. N.

P.S. With the view of economising space I have omitted most of the calculations, which, however, are very simple, and can be verified in a moment by the reader. The equivalents thus obtained by calculation will be found to approximate those procured by experiment, as closely as can be expected in such cases.

I also freely admit that some of the relations above pointed out are more apparent than real; others, I trust, will prove of a more durable and satisfactory description.


Relations between Equivalents[20]

John A. R. Newlands, Chemical News 10, 59-60 (July 30, 1864).

To the Editor of the CHEMICAL NEWS.

SIR,-- In your impression of the 2nd inst. a correspondent, under the name of "Studiosus," has called attention to the existence of a law to the effect "that the atomic weights of the elementary bodies are, with few exceptions, either exactly or very nearly multiples of eight."[21]

Now, in a letter "On Relations among the Equivalents," which was signed with my initials, and inserted in the CHEMICAL NEWS of February 7, 1863, I called attention to the numerical differences between the equivalents of certain allied elements, and showed that such differences were generally multiples of eight, as in the following examples:--[22]

Member of a Group havingOne immediately aboveDifference.
Lowest Equivalent.the Preceding.H=1O=1
Magnesium 24Calcium 40161
Oxygen 16Sulphur 32161
Lithium 7Sodium 23 161
Carbon 12Silicon 28161
Fluorine 19Chlorine 35.516.51.031
Nitrogen 14Phosphorus 31171.062
Lowest Term of Triad.Highest Term of Triad.
Lithium 7Potassium 39322
Magnesium 24Cadmium 112885.5
Molybdenum 96Tungsten 184885.5
Phosphorus 31Antimony 122915.687
Chlorine 35.5Iodine 12791.55.718
Potassium 39Caesium 133945.875
Sulphur 32Tellurium 129976.062
Calcium 40Barium 137976.062

In the last of the above columns the difference is given referred to 16, the equivalent of oxygen, as unity, and it will be seen that, generally speaking, the equivalent of oxygen is the unit of these differences, just as the equivalent of hydrogen, in "Prout's law," is the unit of the atomic weights. Exceptions there are, however, in both cases which render it necessary to take one half or one quarter of the equivalent of oxygen in the one case, and of hydrogen in the other, in order to represent all the numbers obtained as multiples by a whole number of the given standard.

Now, if the law of "Studiosus" had any real existence, the above facts would resolve themselves into particular cases of its application. For if "the atomic weights are multiples of eight," any differences between them must also be divisible by eight. We have here the symbols and the atomic weights of sixty-one elements, placed in their numerical order,[23] and in the third column is the difference between each atomic weight and the one immediately preceding it:--

H1Ca401Ce922.5V1370
Li76Ti5010La920Ta1381
G[24]92Cr52.52.5Di964W18446
B112Mn552.5Mo960Nb19511
C121Fe561Ro[25]1048Au1961
N142Co58.52.5Ru1040Pt1971
O162Ni58.50Pd106.52.5Ir1970
Fl[26]193Cu63.55Ag1081.5Os1992
Na234Y640.5Cd1124Hg2001
Mg241Zn651Sn1186Tl2033
Al27.53.5As7510U1202Pb2074
Si280.5Se79.54.5Sb1222Bi2103
P313Br800.5I1275Th23828
S321Rb855Te1292
Cl35.53.5Sr87.52.5Cs1334
K393.5Zr89.52Ba1374

Now, it will be observed that in all the above differences the number eight occurs but once, and we never meet with a multiple of eight, whereas if the law of "Studiosus" were true the equivalents of the elements, in whatever order they might be placed, should, when not identically the same, differ either by eight or by some multiple of eight in every case.

While upon the subject of "relations among the equivalents," I may observe that the most important of these may be seen at a glance in the following table:--

Triad.
Lowest term.Mean.Highest term.
I.Li 7+17 = Mg 24Zn 65Cd 112
II.B 11Au 196
III.C 12+16 = Si 28Sn 118
IV.N 14+17 = P 31As 75Sb122+88 = Bi 210
V.O 16+16 = S 32Se 79.5Te 129+70 = Os 199
VI.F 19+16.5 = Cl 35.5Br 80I 127
VII.Li 7+16 = Na 23+16 = K 39Rb 85Cs 133+70 = Tl 203
VIII.Li 7+17 = Mg 24+16 = Ca 40Sr 87.5Ba 137+70 = Pb 207
IX.Mo 96V 137W 184
X.Pd 106.5Pt 197

This table is my no means as perfect as it might be[27]; in fact, I have some by me of a more complete character, but as the position to be occupied by the various elements is open to considerable controversy, the above only is given as containing little more than those elementary groups the existence of which is almost universally acknowledged.

I now subjoin a few explanatory remarks on the different groups contained in the above table, the number attached to each group being merely for the purpose of reference.

Group II.-- Boron is here classed with gold, both these elements being triatomic, although the latter is sometimes monatomic.

Group III.-- Silicon and tin stand to each other as the extremities of a triad. Titanium is usually classed along with them, and occupies a position intermediate between silicon and the central term or mean of the triad, which is at present wanting[28]; thus,

(Si 28 + Sn 118)/2 = 73, mean of triad, and
(Si 28 + Mean of triad 73)/2 = 50.5, the eq. of Ti being 50.
Group IV.-- The equivalent of antimony is nearly the mean of those of phosphorus and bismuth; thus,
(31+210)/2 = 120.5, the eq. of Sb being 122.
Group VII.-- The relations which M. Dumas has pointed out between the members of this group are well known; a slight alteration must be made, owing to the atomic weight of caesium having been raised. The relations, then, will be thus:--
Li + K = 2 Na, or in figures, 7 + 39 = 46
Li + 2 K = Rb, " " 7 + 78 = 85
2 Li + 3 K = Cs, " " 14 + 117 = 131
Li + 5 K = Tl, " " 7 + 195 = 202
3 Li + 5 K = 2 Ag, " " 21 + 195 = 216

The equivalent of silver is thus connected with those of the alkali metals. It may also, which amounts to the same thing, be viewed as made up of the equivalents of sodium and rubidium, thus, 23 + 85 = 108. It is likewise nearly the mean between rubidium and caesium, thus, (85+133)/2 = 109.

Group VIII.-- If lithium may be considered as connected with this group as well as with the foregoing (and by some chemists its oxide is viewed as a connecting link between the alkalies and the alkaline earths), we may perform the same calculations in this group that M. Dumas has done in the preceding, thus,--

Li + Ca = 2 Mg, or in figures, 7 + 40 = 47
Li + 2 Ca = Sr " " 7 + 80 = 87
2 Li + 3 Ca = Ba " " 14 + 120 = 134
Li + 5 Ca = Pb " " 7 + 200 = 207

Again, there are two triads in the group of alkali metals, one which has been long known--viz., lithium, sodium, and potassium, and the other, which was pointed out by Mr. C. W. Quin, in the CHEMICAL NEWS of November 9, 1861--viz., potassium, rubidium, and caesium. Potassium is thus the highest term of one triad and the lowest term of another.

In like manner, if we include lithium, we shall have among the metals of the alkaline earths two triads, the first comprising lithium, magnesium, and calcium, and the second calcium, strontium, and barium, calcium standing at the top of one triad and at the bottom of the other.

The element lead occupies a position in relation to the metals of the alkaline earths similar to that filled by thallium in the group of alkali metals. Osmium appears to play a similar part in the sulphur group, and bismuth in the phosphorus group. The analogous term in the chlorine group is not yet known.[29] Thallium, in its physical properties, bears some resemblance to lead, and it frequently happens that similar terms taken from different groups, such as oxygen and nitrogen, or sulphur and phosphorus, bear more physical resemblance to each other than they do to the members of the groups to which, for chemical reasons we are compelled to assign them.

It will be observed that the difference between the equivalents of tellurium and osmium, caesium, and thallium, and barium and lead, respectively, is the same in each case--viz., 70.

Group X.-- Palladium and platinum appear to be the extremities of a triad, the mean of which is unknown.[30]

So frequently are relations to be met with among the equivalents of allied elements, that we may almost predict that the next equivalent determined, that of indium, for instance, will be found to bear a simple relation to those of the group to which it will be assigned.

In conclusion, I may mention that the equivalents I have adopted in this letter were taken from the highly-interesting and important paper by Professor Williamson, lately published in the Journal of the Chemical Society.

I am, &c.

John A. R. Newlands, F.C.S.

Laboratory, 19, Great St. Helens, E. C., July 12.


On Relations Among the Equivalents

John A. R. Newlands, Chemical News 10, 94-95 (Aug. 20, 1864).

To the Editor of the CHEMICAL NEWS.

SIR,-- In addition to the facts stated in my late communication, may I be permitted to observe that if the elements are arranged in the order of their equivalents, calling hydrogen 1, lithium 2, glucinum 3, boron 4, and so on (a separate number[31] being attached to each element having a distinct equivalent of its own, and where two elements happen to have the same equivalent, both being designated by the same number), it will be observed that elements having consecutive numbers frequently either belong to the same group or occupy similar positions in different groups, as in the following examples:--[32]

No.No.No.No.No.
Groupa.N6P13As26Sb40Bi54
"b.O7S14Se27Te42Os50
"c.Fl8Cl15Br28I41----
"d.Na9K16Rb29Cs43Tl52
"e.Mg10Ca17Sr30Ba44Pb53
Here the difference between the number of the lowest member of a group and that immediately above it is 7; in other words, the eighth element starting from a given one is a kind of repetition of the first, like the eighth note of an octave in music.[33] The differences between the numbers of the other members of a group are frequently twice as great; thus in the nitrogen group, between N and P there are 7 elements; between P and As, 13; between As and Sb, 14; and between Sb and Bi, 14.

In conclusion, I may remark that just as we have several examples of the apparent existence of triads, the extremities of which are known, whilst their centres are wanting (such as the metals of the platinum group, which may be conceived to be the extremities of three distinct triads, and perhaps also silver and gold may be related to each other in this manner[34]), so we may look upon certain of the elements, e.g., Mn, Fe, Co, Ni, and Cu, as the centres of triads, the extremes of which are at present unknown, or, perhaps, in some cases only unrecognised.

I am, &c.

John A. R. Newlands, F.C.S.

Laboratory, 19, Great St. Helens, E. C., August 8.


On the Law of Octaves

John A. R. Newlands, Chemical News 12, 83 (Aug. 18, 1865).

To the Editor of the CHEMICAL NEWS.

SIR,-- With your permission, I would again call attention to a fact pointed out in a communication of mine, inserted in the CHEMICAL NEWS for August 20, 1864.

If the elements are arranged in the order of their equivalents, with a few slight transpositions, as in the accompanying table[35], it will be observed that elements belonging to the same group usually appear on the same horizontal line.[36]

No.No.No.No.No.No.No.No.
H[37] 1F 8Cl 15Co & Ni 22Br 29Pd 36I 42Pt & Ir 50
Li 2Na 9K 16Cu 23Rb 30Ag 37Cs 44Tl 51
G 3Mg 10Ca 17Zn 25Sr 31Bd [sic-Cd] 38Ba & V 45Pb 54
Bo[38] 4Al 11Cr 19Y 24Ce & La 33U 40Ta 46Th 56
C 5Si 12Ti 18In 26Zr 32Sn 39W 47Hg 52
N 6P 13Mn 20As 27Di & Mo 34Sb 41Nb 48Bi 55
O 7S 14Fe 21Se 28Ro & Ru 35Te 43Au 49Os 51
(NOTE.-- Where two elements happen to have the same equivalent, both are designated by the same number.)

It will also be seen that the numbers of analogous elements generally differ either by 7 or by some multiple of seven; in other words, members of the same group stand to each other in the same relation as the extremities of one or more octaves in music. Thus, in the nitrogen group, between nitrogen and phosphorus there are 7 elements; between phosphorus and arsenic, 14; between arsenic and antimony, 14; and lastly, between antimony and bismuth, 14 also.[39]

This peculiar relationship I propose to provisionally term the "Law of Octaves."

I am, &c.

John A. R. Newlands, F.C.S.

Laboratory, 19, Great St. Helen's, E.C., August 8, 1865.


Proceedings of Societies [Report on the Law of Octaves]

Chemical News 13, 113 (March 9, 1866).

PROCEEDINGS OF SOCIETIES.
CHEMICAL SOCIETY. Thursday, March 1.

...[40]

Mr. JOHN A. R. NEWLANDS read a paper entitled "The Law of Octaves, and the Causes of Numerical Relations among the Atomic Weights."[41] The author claims the discovery of a law according to which the elements analogous in their properties exhibit peculiar relationships, similar to those subsisting in music between a note and its octave. Starting from the atomic weights on Cannizzarro's [sic] system, the author arranges the known elements in order of succession, beginning with the lowest atomic weight (hydrogen) and ending with thorium (=231.5); placing, however, nickel and cobalt, platinum and iridium, cerium and lanthanum, &c., in positions of absolute equality or in the same line. The fifty-six elements[42] so arranged are said to form the compass of eight octaves, and the author finds that chlorine, bromine, iodine, and fluorine are thus brought into the same line, or occupy corresponding places in his scale. Nitrogen and phosphorus, oxygen and sulphur, &c., are also considered as forming true octaves. The author's supposition will be exemplified in Table II., shown to the meeting, and here subjoined:--

Table II.--Elements arranged in Octaves.[43]
No.No.No.No.No.No.No.No.
H 1F 8Cl 15Co & Ni 22Br 29Pd 36I 42Pt & Ir 50
Li 2Na 9K 16Cu 23Rb 30Ag 37Cs 44Os 51
G 3Mg10Ca 17Zn 24Sr 31Cd 38Ba & V 45Hg 52
Bo 4Al 11Cr 19Y 25Ce & La 33U 40Ta 46Tl 53
C 5Si 12Ti 18In 26Zr 32Sn 39W 47Pb 54
N 6P 13Mn 20As 27Di & Mo 34Sb 41Nb 48Bi 55
O 7S 14Fe 21Se 28Ro & Ru 35Te 43Au 49Th 56

Dr. GLADSTONE[44] made objection on the score of its having been assumed that no elements remain to be discovered. The last few years had brought forth thallium, indium, caesium, and rubidium, and now the finding of one more would throw out the whole system.[45] The speaker believed there was as close an analogy subsisting between the metals named in the last vertical column as in any of the elements standing on the same horizontal line.

Professor G. C. FOSTER humorously inquired of Mr. Newlands whether he had ever examined the elements according to the order of their initial letters? For he believed that any arrangement would present occasional coincidences, but he condemned one which placed so far apart manganese and chromium, or iron from nickel and cobalt.[46]

Mr. NEWLANDS said that he had tried several other schemes before arriving at that now proposed. One founded upon the specific gravity of the elements had altogether failed, and no relation could be worked out of the atomic weights under any other system than that of Cannizzarro.

...


Notes

[1]The terms equivalent, equivalent weight, equivalent number, and atomic weight may be taken as synonymous with combining weight. All of these terms refer to the mass of a given element relative to another element (usually hydrogen = 1). The term atomic weight was in use during Newlands' time; however, many chemists avoided the term in an era when the existence of atoms was a controversial subject. Molar mass is a modern term for essentially the same thing.

As the title makes clear, the primary concern of this paper is to explore relationships between the atomic weights of elements whose properties were already known to be related. As yet Newlands was still dealing with groups of elements, and not yet looking at any relationships among groups. It represents an interesting step toward discovery of a periodic classification of elements.

[2]Jean Baptiste André Dumas (1800-1884; see portrait at the Edgar Fahs Smith collection, University of Pennsylvania) made many contributions to chemistry, principally in the area of organic chemistry. Dumas developed a method for determining the molecular weight of substances which could be turned into vapors. As Newlands mentions, Dumas also made contributions in the area of relations of atomic weights within families of elements [Dumas 1857].

[3]By natural family or group, Newlands means sets of elements which have similar properties. For example, lithium, sodium, and potassium, are all soft metals, react vigorously with water, form salts with one equivalent of chlorine, etc.

[4]These atomic weights are largely based on those of Gerhardt [van Spronsen 1969], with most metals assigned weights roughly half their currently accepted values. Except for Groups I, V, and VII (in Newlands' classification), most of the atomic weights listed here are not those currently recognized. Not until his next paper [Newlands 1864a] did Newlands use atomic weights based on the reforms Cannizzaro urged at the 1860 Karlsruhe Congress. [Cannizzaro 1858] Newlands did not attend Karlsruhe [van Spronsen 1969]. Indeed, during 1860 Newlands was a volunteer in Garibaldi's army fighting in the revolution which would unify Italy. (Newlands' mother was an English woman of Italian descent.)

[5]Newlands' alkali metal group is essentially the same as that recognized today except for its inclusion of thallium. (Thallium was frequently misclassified with the alkali metals; see chapter 12, note 4.) Some modern tables include hydrogen with the alkali metals, despite important differences between hydrogen and the metals.

[6]On the basis of a pattern in atomic weights, Newlands predicts the existence of an unknown member of the alkali metal group. This prediction proved to be incorrect. The element whose atomic weight is 163 is dysprosium (Dy), a rare earth unrelated to the alkali metals. The incorrect prediction is based on an incorrect premise of a regularity in atomic weight intervals.

[7]Newlands' alkaline earths group is essentially the same as that recognized today except for its neglect of beryllium. Today the group also includes radium, not yet discovered at that time.

[8]This does not correspond to any group in the present classification system.

[9]Didymium appears in many lists of elements at this time, but it was later proved to be a mixture of two elements, praseodymium and neodymium. It is found in Newlands' tables below abbreviated as Di.

[10]Newlands classes together elements whose compounds with one atom of oxygen (protoxides) have the same crystal shape as magnesia (MgO). Indeed, the protoxides of most of these elements form cubic crystals. Except for magnesium and cadmium, all of these elements are in the first row of transition metals in the modern periodic table. Note that this is the second group to which Newlands has assigned magnesium.

[11]The halogen group of the modern periodic classification contains these elements as well as astatine, not yet discovered (or rather synthesized) in Newlands' time.

[12]The corresponding group of the modern periodic classification contains these elements as well as polonium, not yet discovered in Newlands' time.

[13]Except for osmium, this group is identical with the nitrogen group in the modern periodic classification.

[14]Today's periodic classification groups carbon, silicon, and tin together with lead (commonly misclassified in Newlands' time; see chapter 12, note 4) and germanium (not yet known). Titanium is not part of this group, but shares with its members the ability to make compounds with two oxygen atoms or four chlorine atoms.

[15]This is not recognized as a group in the modern classification, but rather several elements from adjacent columns of transition metals.

[16]Ruthenium, rhodium, and palladium are three closely related transition metals, analogous to the trio osmium, iridium, and platinum.

[17]These elements do not form a group according to the modern periodic table. They do, however, have a property in common which makes qualitative analytical chemists consider them as a group: their compounds with chlorine are not soluble in water.

[18]Newlands is looking for patterns, a common occupation of scientists, and he seems to have found something. At the same time, he appears to be so sure of the existence of patterns that he offers examples which clearly do not fit the pattern. For example, the difference in the atomic weights he is using for magnesium and cadmium is 44, which is as distant as possible from a multiple of 8. Similarly the example of chlorine and iodine is far from a multiple of 8. It may be argued that most of Newlands' atomic weights were incorrect anyway; however, the reasoning is flawed as well.

[19]This prediction proved to be incorrect on several counts. First, the cited atomic weights of iridium and rhodium are about half their currently accepted values. (See note 4 above.) Furthermore, iridium and rhodium have no analogues of intermediate atomic weight. The element whose atomic weight lies halfway between them is promethium (Pm), a rare earth unrelated to them.

[20]Like the previous paper, this one deals with relationships between the atomic weights of elements whose properties were already known to be related. Newlands begins to put together a classification system from groups of triads and extended triads [van Spronsen 1969]. It represents a further step toward discovery of a periodic classification of elements, but it certainly does not describe periodic trends.

[21]"Studiosus," an anonymous correspondent to the Chemical News, was also looking for patterns among atomic weights. Unfortunately, the one he "found" does not exist. Studiosus' claim seems to be similar to a speculation by Prout (chapter 10) that all elements have atomic weights which are multiples of those of hydrogen, suggesting that hydrogen is the "primary matter" from which all other material is made. A few paragraphs later, Newlands refers to Prout's hypothesis as a law. He also suggests an analogous relationship, that the differences in atomic weight of allied elements may be multiples of 8, but he notes correctly that the differences between succeeding elements is seldom 8 (as it would have to be if Studiosus were correct).

[22]Most of the examples are the same as in the previous paper, but many of the atomic weights are different. Over the previous year, Newlands had switched to using atomic weights based on Cannizzaro's system [Cannizzaro 1858], a fact he mentions only in the article's last sentence.

[23]Newlands later claimed that this table "was the first ever published" of all the known elements in the order of atomic weight [Newlands 1884]. The listing of elements by atomic weight is so common today that the claim sounds incredible. Even though it is not strictly correct (for example, John Hall Gladstone had published such an arrangement, albeit with many unreliable atomic weights, in 1853 [Gladstone 1853], and even Dalton's incomplete list of unreliable weights was in numerical order [Dalton 1808]), the arrangement was sufficiently unusual even in 1875 that Newlands published a note extolling its advantages [Newlands 1875].

[24]G stands for glucinum, an alternative name for beryllium (Be). Neither the atomic weight nor the characteristic valence of beryllium was firmly established at this time. (See chapter 13, note 38 for details.) Newlands uses the correct atomic weight here, and classifies beryllium correctly (with magnesium, etc.) below [Newlands 1865aNewlands 1865a]. But not until many years later did he claim that his classification could be used to decide between competing values of the atomic weight [Newlands 1878b]. In fact, in the absence of such a statement, the natural assumption for the reader is to take the atomic weight presented here as a given which led to the subsequent correct classification of beryllium. By contrast, Mendeleev made an explicit link between his system of classification and his choice of atomic weight for beryllium [Mendeleev 1872].

[25]The standard symbol for rhodium is Rh.

[26]The standard symbol for fluorine is F.

[27]This table exhibits two striking features of the periodic classifications that emerged in the 1860s. First, most of the table shows a recurrence of analogous elements with increasing atomic weight. (Indeed, rows I through VIII and the columns except the last put elements in order of atomic weight.) Second, there is a gap in row III for an unknown element. The table has many flaws as well, though. Rows IX and X ruin the order of atomic weight; row X leaves room for an unknown element that does not exist. Most of the last column is misclassified. Lithium and magnesium appear in multiple locations. Osmium and lead, misclassified in the previous paper, are misclassified differently here. And as Newlands says, the table is not a complete classification, but shows only 37 of the approximately 60 elements known at that time.

[28]Newlands predicts the existence of an unknown element of atomic weight 73. This prediction proved to be correct. Clemens Winkler discovered germanium (atomic weight 72.6) in 1886. [Winkler 1886] It is an analogue of silicon and tin, and also shares a valence of 4 with titanium. There are several respects in which this prediction is better than others Newlands made--aside from the fact that it turned out to be right. First, this prediction is based on more than one piece of evidence, more than one triad. An individual triad is a somewhat flimsy basis for prediction, but two support each other. Second, the prediction concerns lighter atoms which were better known at the time. If the list of elements can be compared to a jigsaw puzzle, it is certainly easier to note the size of a piece missing from an area of the puzzle with relatively few gaps and whose patterns are fairly well known than to predict a piece missing from an area where gaps abound and patterns are little more than speculations.

Newlands reasserted his priority in predicting this element after Mendeleev's prediction of another element (gallium) proved correct but before germanium was actually discovered. [Newlands 1878b, 1884] I say reasserted because the "law of octaves" in the form in which Newlands later introduced [Newlands 1865a, Chemical News 1866] it was inconsistent with this prediction.

[29]This prediction of a heavy halogen turned out to be correct, borne out by the synthesis of astatine in 1940. How much credit Newlands deserves for the prediction is highly debatable. Newlands argues that there must be a heavy member of the chlorine group just as there are heavy members of the alkali metals, alkaline earths, phosphorus, and sulfur groups. His analogies were faulty, however, for lead is not an alkaline earth (but belongs to the silicon group); thallium is not an alkali metal (but belongs to the aluminum group); and osmium does not belong to the sulfur group (but is a transition metal); only bismuth belongs to the group to which he assigns it. Newlands' prediction obviously would carry more weight if he had based it on group assignments we recognize today as correct. It would even have been more credible had it been more consistent. He recognizes that the chlorine group lacked a heavy member he believed the alkali metals, alkaline earths, phosphorus, and sulfur groups had; why did he not note that (according to his classification) the aluminum and silicon groups also lacked such heavy members?

[30]This prediction of an analogue of palladium and platinum with intermediate atomic weight proved to be incorrect. The element whose atomic weight lies halfway between them is samarium (Sm), a rare earth unrelated to them.

[31]The introduction of an ordinal number for each element represented a significant step. Modern readers may be inclined to see atomic numbers in these ordinal numbers. Atomic numbers, however, are much more fundamental entities, for they correspond to a physical quantity (namely the number of positive charges in the nucleus of the atom). Newlands' contemporaries regarded atomic weight as more fundamental than these order numbers, and rightly so, for the atomic weight corresponded to a measurable physical quantity. Yet the ordinal numbers correlated with the atomic weights, a point Newlands made explicit in later publications [Newlands 1865b, 1878a]. Newlands used these numbers as surrogates for atomic weights--surrogates that showed more regularity than atomic weights.

[32]This table can be regarded as only a fragment of a classification system, for it includes only 24 elements of the roughly 60 known to Newlands--fewer even than the table in his previous paper. Furthermore, the last column contributes nothing of value to the table, for most of the elements are placed incorrectly with respect to both atomic weight order and chemical group. Still, there is undeniably a periodic pattern here.

[33]The eighth tone in any major or minor musical scale is a repetition of the original tone one octave higher (or in physical terms, at double the frequency). Newlands was so taken with the fact that the chemical properties seemed to repeat with the eighth element that he subsequently dubbed this observation the "law of octaves." Several writers since Newlands' day have commented that this imaginative analogy may have prevented his more prosaic contemporaries from taking his work as seriously as it deserved.

[34]Newlands refers to the prediction he made near the end of his previous paper of an analogue of palladium and platinum with intermediate atomic weight, suggesting here that an analogue of silver and gold also exists with intermediate atomic weight. Neither speculation proved to be correct.

[35]The table includes all the elements known at the time, so it can be regarded as an attempt at a system of classification of the elements.

[36]Clearly Newlands has enunciated the principle of periodicity (repetition of chemical properties in a series of elements arranged by atomic weight), however imperfectly his table displays it. And the table is far from perfect. For instance, it requires several inversions in atomic weight order (Cr-Ti, Zn-Y, Ce & La-Zr, U-Sn, Te-I, and most of the last column), only one of which is retained in the current periodic table. Some of the inversions, particularly in the last row, not only place elements out of their numerical order but remove them from the chemical groups to which they should belong.

[37]Newlands considered hydrogen to be analogous to the halogens as seen from this classification and stated explicitly in a later publication [Newlands 1872, 1896].

[38]The standard abbreviation for boron is B.

[39]Consideration of the table alone can give rise to an impression Newlands apparently did not intend, namely that all elements in the same row constitute a group. Note that Newlands does not say that the horizontal lines of his table constitute groups of analogous elements, but rather that such groups are found on horizontal lines (perhaps with other elements). This paragraph and the one immediately preceding the table are carefully worded. Clearly Newlands counts nitrogen, phosphorus, arsenic, antimony, and bismuth in the same group; he does not say manganese, didymium, molybdenum, and niobium are part of that group. Still, Newlands must bear some responsibility for misunderstandings. What is a clear inference upon close reading is not clear upon cursory reading--or upon focusing on the table; furthermore, since the table is the most prominent part of this communication special care must be taken to prevent it from conveying a mistaken impression.

[40]The following is an account of a paper Newlands read on the law of octaves before the Chemical Society. It was written by a reporter whose name I do not have. I have omitted the parts of the report which do not deal with Newlands' presentation.

[41]The Chemical Society did not publish the paper in its Journal. In 1873 Newlands, attempting to assert his priority as discoverer of the periodic law, asked that the Chemical Society publish a note referring to his 1866 presentation. William Odling, President of the Society, said that the reason why Mr. Newlands' paper on this subject in 1866 had not been published by the Society was that they had made it a rule not to publish papers of a purely theoretical nature, since they were likely to lead to correspondence of a controversial character. [Newlands 1873] Newlands went on to mount quite a campaign asserting his priority in the matter, publishing a booklet in 1884 which collected his work on the subject [Newlands 1884].

Interestingly enough, Odling had also worked in the area of classification of elements and relationships among their atomic weights. He published a classification in 1864 [Odling 1864] which was similar in several ways to that of Mendeleev in 1869 [Mendeleev 1869]. Although Odling never claimed priority for the periodic law, several historians of chemistry number him among its discoverers [Cassebaum & Kauffman 1970, van Spronsen 1969].

[42]Actually, there are 62 elements in the table, with 6 pairs assigned the same order number.

[43]This table contains fewer inversions in atomic weight order than the 1865 paper [Newlands 1865a]: yttrium and zinc as well as all of the final column are in atomic weight order, at least according to the atomic weights available to Newlands. This table places yttrium and thallium in the same group as boron and aluminum; the modern table does not place them in the same group, but recognizes that both are trivalent. This table places zinc and mercury together, correctly; it places them in the same group as beryllium and magnesium, with which they at least share bivalence. Finally, this table places lead, correctly, in the same group with carbon and silicon.

[44]John Hall Gladstone had himself published on the subject of numerical relations among equivalent weights and classification of the elements [van Spronsen 1969; see Gladstone 1853].

[45]Newlands' attempt to answer this criticism [Newlands, 1866] was quite unsatisfactory. He maintained, "The fact that such a simple relation exists now, affords a strong presumptive proof that it will always continue to exist, even should hundreds of new elements be discovered." Perhaps. But by not allowing for blank spaces, Newlands made his system both less accurate and less powerful than it could have been. Less accurate, for he had to place some elements out of order in order to fill in gaps left by elements then unknown (for example, putting chromium before titanium, in the place now occupied by scandium). Less powerful because the system, unlike that of Mendeleev or even his earlier attempts, made no attempt to be predictive, to assist in the discovery of new elements. Newlands went on to add that if new elements are discovered, "the difference in the numbers of analogous elements might, in that case, be altered from 7, or a multiple of 7, to 8, 9, 10, 20, or any conceivable figure, the existence of a simple relation among the numbers of analogous elements would be none the less evident." This statement implies that new elements will require new groups, which proved true for argon but not for most elements discovered after this time. Finally, Newlands offered, "As a proof, however, that new discoveries are not very likely to destroy such relationship, I may mention that when the existence of the 'law of octaves' was first pointed out (Chemical News, August 20, 1864 [Newlands 1865b]), the difference between the numbers of P and As was 13 instead of 14, as between As and Sb, and also between Sb and Bi. Since then, by the determination of the atomic weight of indium, the difference of the numbers of P and As has been made to be 14, as in the other cases adduced." His argument here would be faulty, even if he had placed indium correctly. The insertion of indium between phosphorus and arsenic caused the latter elements to fall into octaves; it should be obvious that insertion of another element in a similar way would disrupt whatever octaves already existed.

[46]Foster recognized the similarities among metals we now recognize as members of the first transition series, and Gladstone made a similar comment about the heavy metals in the last vertical column. Both of these critics recognized similarities among elements which did not fall into horizontal lines. Newlands quickly responded to this criticism in print [Newlands 1866], noting, "The rule … was expressed as follows:-- 'The numbers of analogous elements, when not consecutive, differ by 7, or by some multiple of 7.' The clause 'when not consecutive' was introduced for the purpose of embracing certain analogous elements whose atomic weights are consecutive." In Newlands' defense, it must be pointed out that he mentioned explicitly the possibility of analogous elements having consecutive numbers in a paper reproduced above [Newlands 1864b] as well as in a later 1864 note [Newlands 1864c]. In defense of Newlands' critics, the paper in which Newlands introduced the term "law of octaves" made no mention of groups of consecutive elements [Newlands 1865a], and even the earlier 1864 paper gave examples only of groups that contain intervening elements.

References


Back to the top of the table of contents of Elements and Atoms.
Back to the top of Classic Chemistry.