Atomic weight of Be ?

[ERIC SCERRI]

I have been looking into the early use of atomic weights by the discoverers of the periodic table.

Van Spronsen’s 1969 classic book has a rather good table that displays the atomic weights of many chemists of the day. (pp. 46-49)

I am intrigued by a single entry, that of a value of 9 for beryllium under the column headed De Chancourtois, while Cannizzaro, who is widely credited with having established a set of definitive atomic weights, omits any value whatsoever for Be.  Berzelius thinks it is 53!

The question of the atomic weight of beryllium presented a genuine problem since it was not clear whether it was di- or tri-valent.  

I have a section on this in my 2007 book on the periodic table (pp. 127-8) and again on pp. 141-2 in the second edition of 2020.  

The issue was finally sorted out by Mendeleev, who opted for di-valency, which meant that beryllium was assigned an atomic weight of 9.4.

But De Chancourois’ spiral periodic system appeared a full 7 years before Mendeleev’s first published table of 1869.

Where could De Chancourtois have obtained more or less the same value so early?

Regards

Eric

[johnmarks9]

Alexandre Émil Béguyer de Chancourtois (what a name!) was, of course, a geologist and probably familiar with beryl (Be valency +2) and maybe other beryllium compounds. 

A search of Comptes Rendus and other early 19th century geological, crystallographic or mineralogy journals may be a source worth examining to answer your question.

Regards,

John Marks

[ERIC SCERRI]

Thank you John,

As I tried to explain, the valency of Be was generally taken to be 3, not 2, before Mendeleev’s intervention.

This led to an incorrect atomic weight.

Eric Scerri

-- 
You received this message because you are subscribed to the Google Groups "Periodic table mailing list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to PT-L+uns...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/PT-L/34b04cbe-ca5b-4970-88f0-82d638d46682n%40googlegroups.com.

ERIC SCERRI PhD

UCLA

Website: http://www.ericscerri.com
PhilPeople https://philpeople.org/profiles/eric-scerri
Google Scholar https://scholar.google.com/citations?hl=en&user=8w1T5bEAAAAJ
ResearchGate https://www.researchgate.net/profile/Eric-Scerri

Editor-in-Chief of Foundations of Chemistry https://link.springer.com/journal/10698

 

[Mark Leach]

Hi Eric & All,

Interesting. I had not clocked that Cannizzaro had omitted Be:

https://www.meta-synthesis.com/webbook/35_pt/pt_database.php?PT_id=1047

Or lithium (or helium… which, to be fair, had not been discovered…)

I see from the attached table, that Berzelius had originally said 10.9 which he then modified to 53 (a big jump!).

Perhaps De Chancourtois used the original 1815 Berzelius logic?

Beryl has a complex stoichiometry: Be3Al2(SiO3)6 so it would not be surprising if this gave an erroneous weight for Be. After all, Daubeny thought Al had a weight of 10:

https://www.meta-synthesis.com/webbook/35_pt/pt_database.php?PT_id=337

Mark

Mark Leach

meta-synthesis

--
You received this message because you are subscribed to the Google Groups "Periodic table mailing list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to PT-L+uns...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/PT-L/6A5D112C-DC47-4BBE-92F9-931069253A40%40g.ucla.edu.
<PastedGraphic-1.pdf>

ERIC SCERRI PhD

UCLA

Website: http://www.ericscerri.com
PhilPeople https://philpeople.org/profiles/eric-scerri
Google Scholar https://scholar.google.com/citations?hl=en&user=8w1T5bEAAAAJ
ResearchGate https://www.researchgate.net/profile/Eric-Scerri

Editor-in-Chief of Foundations of Chemistry https://link.springer.com/journal/10698

 

--
You received this message because you are subscribed to the Google Groups "Periodic table mailing list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to PT-L+uns...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/PT-L/6A5D112C-DC47-4BBE-92F9-931069253A40%40g.ucla.edu.

[René]

Hi Eric

This is an interesting topic of which I knew little.

On Berzelius’ value of 53, Parsons (1904, p. 721) wrote that:

"The first determinations of the equivalent of beryllium were made by Berzelius in 1815, and consisted of a single analysis each of an undoubtedly impure hydrous sulphate and chloride, both of which were probably also basic in character. The results are, therefore, of no interest in a discussion of the atomic weight of this element."

Berzelius also favoured Be2O3 as the formula of the oxide which couldn’t have helped.

Parsons (1904, p. 723) went on to list the following determinations of the atomic weight of beryllium:

                              Ratio determined   Mean O = 16

============================================================

1842 Awdejew               BeO : BaS04        9.34

1854 Weeren               BeO : BaS04        9.27

1855 Debray               BeO : 4C02         9.34

------------------------------------------------------------

1869 Klatzo               BeO : BaSO         9.28

1880 Nilson and Petterson  BeS04.4H20 : BeO   9.104

1891 Krüss and Moraht      BeS04.4H20 : BeO   9.05

Evidently De Chancourtois relied on one or more of the first three, all of which predate his 1862 Telluric Screw.

He was in good company since Newlands (1864) and Odling (1864) also showed Be as 9, again predating Mendeleev’s first periodic system of 1869.

Looking closer at De Chancourtois’ Teulluric Screw it shows GlO (i.e. BeO)—in the column to the left—

as the basis for the atomic weight of 9 (rounded down): https://www.meta-synthesis.com/webbook/35_pt/pt_database.php?PT_id=7

That said, as you know, since Be appeared to behave like Al there was an expectation by others that beryllium oxide ought to be Be2O3 rather than BeO, and that its atomic weight would therefore be ~13.5.

There is some discussion on the supposed experimental basis for this higher atomic weight in Nature (1880, pp. 57–58).

As you noted, you also discuss this at pp. 141–142 of the 2nd ed of your Red Book.

De Chancourtois and aluminium

All that said, I don’t understand what you wrote about De Chancourtois grouping Al with the alkali metals (p. 81):

One final comment should perhaps be made about De Chancourtois. His lack of

chemical knowledge may have been a hindrance in some cases, and conversely, his

emphasis on geological factors may have misled him in the development of the periodic

system. For example, he stated that the isomorphism between feldspars and pyroxenes

had been the starting point of his system. The element aluminum appears to function

analogously to the alkali metals, a fact that does not necessarily indicate that aluminum

should be grouped together with alkali metals such as sodium and potassium. But this is

precisely what De Chancourtois did in his system. In fact, he even changed the atomic

weight, or characteristic weight, as he termed it, in the case of aluminum to make it fall

neatly into line with the alkali metals. Had he known more chemistry, he might not have

taken this unjustified step.

On what basis did you conclude that De Chancourtois aligned Al with the alkali metals, and that he changed the atomic weight of Al, to achieve this?

In his Telluric Screw, Al was aligned under B and over Si-Ni-As-La-Pd. Immediately to the left are Li-Na-K-Mn-Rb-Ru-Th. Al thus does not fall into line with the alkali metals.

De Chancourtois also showed Al with an atomic weight of 27, which is close to the accepted value (26.98).

René

• Nature 1880, The atomic weight of beryllium, 23, https://doi.org/10.1038/023057b0
• Newlands JAR 1864, Chemical News, 10, 59-60
• Parsons CL 1904, A revision of the atomic weight of beryllium, J. Am. Chem. Soc., 26, 7, 721–740
• Odling W 1864, On the proportional numbers of the elements," Quarterly Journal of Science, 1, 642-648

[ERIC SCERRI]

Thanks for all this information, René.

On Jan 22, 2026, at 8:40 PM, René <re...@iinet.net.au> wrote:

Hi Eric

This is an interesting topic of which I knew little.

On Berzelius’ value of 53, Parsons (1904, p. 721) wrote that:

"The first determinations of the equivalent of beryllium were made by Berzelius in 1815, and consisted of a single analysis each of an undoubtedly impure hydrous sulphate and chloride, both of which were probably also basic in character. The results are, therefore, of no interest in a discussion of the atomic weight of this element."

Berzelius also favoured Be2O3 as the formula of the oxide which couldn’t have helped.

Parsons (1904, p. 723) went on to list the following determinations of the atomic weight of beryllium:

                              Ratio determined   Mean O = 16

============================================================

1842 Awdejew               BeO : BaS04        9.34

1854 Weeren               BeO : BaS04        9.27

1855 Debray               BeO : 4C02         9.34

------------------------------------------------------------

1869 Klatzo               BeO : BaSO         9.28

1880 Nilson and Petterson  BeS04.4H20 : BeO   9.104

1891 Krüss and Moraht      BeS04.4H20 : BeO   9.05

This is especially useful.  

I would need to look at De C’s original articles to see where he obtained his atomic weights.  I don’t recall whether I looked at them when I put together the book more than 20 years ago now!

Evidently De Chancourtois relied on one or more of the first three, all of which predate his 1862 Telluric Screw.

He was in good company since Newlands (1864) and Odling (1864) also showed Be as 9, again predating Mendeleev’s first periodic system of 1869.

A good point. 

Looking closer at De Chancourtois’ Teulluric Screw it shows GlO (i.e. BeO)—in the column to the left—

as the basis for the atomic weight of 9 (rounded down): https://www.meta-synthesis.com/webbook/35_pt/pt_database.php?PT_id=7

That said, as you know, since Be appeared to behave like Al there was an expectation by others that beryllium oxide ought to be Be2O3 rather than BeO, and that its atomic weight would therefore be ~13.5.

There is some discussion on the supposed experimental basis for this higher atomic weight in Nature (1880, pp. 57–58).

As you noted, you also discuss this at pp. 141–142 of the 2nd ed of your Red Book.

De Chancourtois and aluminium

All that said, I don’t understand what you wrote about De Chancourtois grouping Al with the alkali metals (p. 81):

One final comment should perhaps be made about De Chancourtois. His lack of

chemical knowledge may have been a hindrance in some cases, and conversely, his

emphasis on geological factors may have misled him in the development of the periodic

system. For example, he stated that the isomorphism between feldspars and pyroxenes

had been the starting point of his system. The element aluminum appears to function

analogously to the alkali metals, a fact that does not necessarily indicate that aluminum

should be grouped together with alkali metals such as sodium and potassium. But this is

precisely what De Chancourtois did in his system. In fact, he even changed the atomic

weight, or characteristic weight, as he termed it, in the case of aluminum to make it fall

neatly into line with the alkali metals. Had he known more chemistry, he might not have

taken this unjustified step.

On what basis did you conclude that De Chancourtois aligned Al with the alkali metals, and that he changed the atomic weight of Al, to achieve this?

Sorry I don’t remember why I wrote that.  It would take me too far away from my primary concern to have to dig into this.  As I explained to Mark, I am trying to put together a lecture for the 100th anniversary of Cannizaro’s death in 1826.

Here’s something that may be of interest to you and others here.

Imagine trying to form the beginnings of a periodic table based on Berzelius’ list of atomic weights of 1845.

You arrange the elements in order of increasing atomic weights and cut the sequence in order to reflect chemical similarities.  Here’s how it would look.

It works for groups 15, 16 and 17 (in modern group numbering)

but it fails for the other trhee potential groups as shown below (shown by asteriskes).

---------------------------------------------------------------------------------------------------------------

Now do it using Cannizaro’s atomic weights of 1860.

Of the 7 groups that are formed below, only one of them (Li and Mg) is incorrect as shown by the asterix.  

To me this shows why putting a resonable periodic table together was not possible before Cannizaro published his 1860 values.

Of course there may have been a set of atomic weights after Berzelius 1845 and before Cannizaro that could have done a good job but I would need to look into it further.  Maybe Debray?  Do you know if he published an extensive list like Berzelius and Cannizaro did?  

The question of the atomic weight of beryllium presented a genuine problem since it was not clear whether it was di- or tri-valent.  

I have a section on this in my 2007 book on the periodic table (pp. 127-8) and again on pp. 141-2 in the second edition of 2020.  

The issue was finally sorted out by Mendeleev, who opted for di-valency, which meant that beryllium was assigned an atomic weight of 9.4.

Your data from Awdejew, Weeren etc. suggests that I gave Mendeleev too much credit for arriving at a value of about 9.  

But De Chancourois’ spiral periodic system appeared a full 7 years before Mendeleev’s first published table of 1869.

Where could De Chancourtois have obtained more or less the same value so early?

Regards

Eric

[René]

Thanks Eric for sharing your thoughts on what seems to be an under-lit corner of periodic table history.

Here’s something that may be of interest to you and others here.

[trim]

Now do it using Cannizaro’s atomic weights of 1860.

Of the 7 groups that are formed below, only one of them (Li and Mg) is incorrect as shown by the asterix.

To me this shows why putting a resonable periodic table together was not possible before Cannizaro published his 1860 values.

It seems to me that you’re giving too much credit to Cannizzaro?

In his letter of 1858 he included atomic weight values for H-B-C-N-O and Na-Mg-Al-Si-P-S-Cl but not, as far as I can see, for Li-Be-F, resulting in:

H  B  C  N  O

Na Mg Al Si P S Cl

Of the five groups, 4 are incorrect and there are 2 orphans. His values for P and S are the same (32). 

The sequence of a pentad followed by a heptad lacks the regular segmentation needed for periodicity to become visually apparent. Then again, Cannizzaro wasn't looking for periodicity; he was concerned with getting atomic weights right.

Van Spronsen does however list Cannizzaro values for Li-Be-Fi which were presumably inferred by him (van Spronsen).

Notably, van Spronsen does not give a source aside from Cannizzaro (1860) which is the 1858 letter reproduced as a pamphlet and circulated at the Karlsruhe congress in 1860. I see van Spronsen also listed a value of 31 for P, which is not consistent with Cannizzaro’s value of 32. For that matter, van Spronsen listed 51 values whereas Cannizzaro gave values for only 30 elements.

In contrast, De Chancourtois (1862) included values for H-Li-Be-B-C-N-O-F and Na-Mg-Al-Si-P-S-Cl, ostensibly resulting in:

H

Li Be B  C  N O F

Na Mg Al Si P S Cl

This is indeed the start of a reasonable periodic table.

In this light it may be more reasonable to say that:

• Cannizzaro laid the foundation for putting a reasonable periodic table together;
• De Chancourtois paved the way; and
• Mendeleev seized the day.

regards, René

[ERIC SCERRI]

On Jan 24, 2026, at 9:29 PM, René <re...@iinet.net.au> wrote:

Thanks Eric for sharing your thoughts on what seems to be an under-lit corner of periodic table history.

Here’s something that may be of interest to you and others here.

[trim]

Now do it using Cannizaro’s atomic weights of 1860.

<Screenshot 2026-01-22 at 9.23.00 PM.png>

Of the 7 groups that are formed below, only one of them (Li and Mg) is incorrect as shown by the asterix.

To me this shows why putting a resonable periodic table together was not possible before Cannizaro published his 1860 values.

It seems to me that you’re giving too much credit to Cannizzaro?

In his letter of 1858 he included atomic weight values for H-B-C-N-O and Na-Mg-Al-Si-P-S-Cl but not, as far as I can see, for Li-Be-F, resulting in:

H  B  C  N  O

Na Mg Al Si P S Cl

Of the five groups, 4 are incorrect and there are 2 orphans. His values for P and S are the same (32). 

The sequence of a pentad followed by a heptad lacks the regular segmentation needed for periodicity to become visually apparent. Then again, Cannizzaro wasn't looking for periodicity; he was concerned with getting atomic weights right.

Van Spronsen does however list Cannizzaro values for Li-Be-Fi which were presumably inferred by him (van Spronsen).

Notably, van Spronsen does not give a source aside from Cannizzaro (1860) which is the 1858 letter reproduced as a pamphlet and circulated at the Karlsruhe congress in 1860. I see van Spronsen also listed a value of 31 for P, which is not consistent with Cannizzaro’s value of 32. For that matter, van Spronsen listed 51 values whereas Cannizzaro gave values for only 30 elements.

Yes, I’ve also been looking at Van Spronsen again.  I agree that he does not give any citations for Cannizzaro’s atomic weights.

As Mark Leach pointed out he list Ti as 56 in his Sunto, whereas Van Spronsen reports 50, rather mysteriously.  

Even more puzzling, Van Spronsen gives 60 references in his chapter 3 that contains the table that I shared with atomic weights from vaerious authors, but not one single citation to Cannizzaro !

This is rather disappointing.  

Does anybody know the source of the column headed Cannizzaro 1860 on p. 47 of Van Spronsen?

I am in the process of also asking several Italian colleagues.

Best

Eric

In contrast, De Chancourtois (1862) included values for H-Li-Be-B-C-N-O-F and Na-Mg-Al-Si-P-S-Cl, ostensibly resulting in:

H

Li Be B  C  N O F

Na Mg Al Si P S Cl

This is indeed the start of a reasonable periodic table.

In this light it may be more reasonable to say that:

• Cannizzaro laid the foundation for putting a reasonable periodic table together;
• De Chancourtois paved the way; and
• Mendeleev seized the day.

regards, René