418

DOC.

14

CORRECTION TO EINSTEIN 1906A

Published

in Annalen

der

Physik

34

(1911):

591-592. Dated

Zurich, January

1911,

received

21

January 1911, published

9

March

1911.

[1]Einstein

1906a,

which

is

the

Annalen

version of Einstein's

dissertation,

Einstein

1905j (Vol.

2,

Doc.

15).

For

a comprehensive

historical

analysis

of Einstein's

dissertation,

of

the

corrections

to

it,

and of

the

related

experimental

research,

see

Vol.

2,

the editorial

note,

"Einstein's Disserta-

tion

on

the Determination of Molecular

Dimensions,"

pp.

170-182.

[2]Jacques

Bancelin,

a

student of Jean

Perrin,

performed experiments

in order

to test

Ein-

stein's relation between the

coefficient

of

viscosity

of

a

liquid

with and without

suspended

molecules

(k*

and

k,

respectively):

k*

=

k(1+Qp), where

Q

is

a

constant which Einstein

origi-

nally

determined

to be

1

(see

Einstein

1905j

[Vol.

2,

Doc.

15], p.

17),

and

cp

is

the fraction of

the

volume

occupied

by

the solute molecules. Bancelin

performed

these

experiments on

Perrin's

initiative,

after Einstein had drawn Perrin's attention

to

the method

developed

in his

disserta-

tion

(see

Einstein

to

Jean

Perrin,

11

November

1909).

Bancelin

reported

values for

Q

of

ca.

3.8

or

3.9 to

Perrin

as

well

as

to

Einstein

(for

evidence of

these

reports, see

Einstein

to

Jakob

Laub,

28

December

1910,

and Einstein

to

Jean

Perrin,

12 January 1911).

[3]See

Einstein

to

Ludwig Hopf,

27

December

1910.

Prior

to

asking Hopf,

Einstein had made

an

unsuccessful

attempt

to

find the

error

himself;

see

Einstein

1905j (Vol. 2,

Doc.

15),

note

14.

[4]See

Einstein

1905j (Vol. 2,

Doc.

15),

p.

12.

[5]See

Einstein

1905j (Vol. 2,

Doc.

15),

p.

13.

[6]After

discovering

the

error

in his calculations,

Einstein

suggested

to

Perrin that

an experi-

mental

error

must

be

responsible

for the

remaining discrepancy

between the

experimental

value

Q

=

3.9

and the theoretical value

Q

=

2.5; see

Einstein

to

Jean

Perrin,

11

November

1909.

By

the end of

May 1911,

Bancelin's

experiments

had indeed

provided

him with

a

value for

Q

that

was

closer

to

the theoretical value

(Q

=

2.9);

see

Bancelin

1911a,

1911b.

Einstein

apparently

also

intended

to

continue work

on

the

phenomenon

of

viscosity; see

Einstein

to

Heinrich

Zangger,

20

September

1911.

[7]See

Einstein

1905j (Vol.

2,

Doc.

15), p.

18.

[8]The

value for N

given by

Einstein

is

found

in

Einstein

1906c

(Vol.

2,

Doc.

33),

p.

306.