214

DOCS.

295,

296 OCTOBER

1911

considered

in

the

case

of

photospheric lines,

but

displacement

due

to diffusion,

as

you

explained

it,[6]

certainly

can.

The

question

would

now

be,

whether

this effect

can give

the

observed value

quantitatively;

I have not

yet

had

time to

consider whether

a

definite

answer

to this is

possible.

With

best

regards,

I

remain

yours truly,

A.

Einstein

296. To Michele

Besso

Prague, 21

October

1911

Dear

Michele,

Don't

be

angry

because of

my long

silence. This time I do not feel

guilty,

because

I

really

did

not

have

a

single

free

moment. I

was

away 3

weeks,

first in

Karlsruhe

at

the

Naturforschervers.,[1]

and

then

in

Zurich,

where

I

had

to

give

8

lectures

at

the vacation

course.[2]

Add

to this

the

many

shoptalks

and

personal obligations! But, now-once

the

witches'

sabbath

in

Brussels[3] will also be

over-I

will be

my own

master

again, except

for

my courses.

I

haven't been

able

to

give

much

thought

to

your

letters,

but

from time to time I

have

been

looking

into them.

The

formula[4]

a

=

82

3h

2n

expresses

the

independence

of

a

from T for

alloys

but,

under

plausible

assumptions,

a

proportionality

with

the

square

of

the admixture.[5]

Further,

it

is

also

not

very

plausible

to set

X2_1T

for

pure

metals and

higher temperatures.

A

simple

temperature

dependence

of n-if

there

is

one-ought

not to

be

assumed without

compelling

reasons

for

it.

It

seems

to

me

most probable

that the

expression

for

a

is proportional to

X

and

that

it otherwise

does

not

contain

any

temperature

variable.

Only

so

can

the

empirical

results

be

comprehended

in

a

natural

way; n

would

then

have to be

without influence

on

the

resistance.

There

is nothing particularly

surprising

in

that.

I believe

that

one

should

investigate

whether it

is

possible

to arrive at this view in

a

more

or

less

unforced

manner.

Sommerfeld formulated

his

hypothesis

about the

collision time

without

any

theory.[7]

But

one can

advance

really weighty

reasons

for

this

hypothesis directly

from

the radiation

formula. It

is not

compatible

with

our

mechanics;

it's

pointless to

rack one's brain

over

this matter.

The radiation formula

shows

that

electrically

charged

elementary

structures

that

collide with low

speed

(energy)

do not

emit

short-wave

radiation. If

one

adheres

more

or

less to

Maxwell's

mechanism

for

emission,

then

it follows

that the

higher

terms

must be

missing

in

the Fourier

expansion

of

the

collision

accelerations,

which

is