592

DOCS.

565,

566

JUNE

1918

which

is

critical

now

to the

relative

quantities of

both

previously equal

bodies.

I

know

very

well

that

Weyl

does

not

recognize

this.

He

would

say,

clocks

and

measuring

rods would have to

appear merely

as

solutions; they

do not

occur

in

the

foundations of

the

theory.[7]

I

find, though:

if

ds,

measured

with

a

clock

(or

a

measuring

rod), is

an

entity

independent

of

the

clock’s

prehistory, design,

and

material

composition,

then

this invariant must

as

such also

play

a

very

funda-

mental

role in

the

theory.

But

if real

natural

processes

were

not like

this,

there

would be

no

spectral

lines and

no

well-defined chemical elements.

Moreover,

for

Weyl

there

is

one

world line of

a

preferred

nature,

the

geodesic.

On

the other

hand,

in nature

there

is

one

type

of

preferred

world

line,

the

trajec-

tories of

electrically uncharged

point

masses.

It

is hardly

conceivable otherwise

than that both

are

identical. The

^’s

now

enter

the

equation

of

this

geodesic

line

as

moving forces,

and this in

a very

dubious

way.[8]

You

can figure

this

out

yourself.

All in

all: I

am

fully

and

firmly

convinced

that this

theory

does not

correspond

to

reality. Formally

as well,

it

appears

unsatisfactory

to

me

that the

angle

be

regarded

as

invariant to

displacement,

but

not

the

unit of

length.-[9]

I

was

very pleased

that

you

are no

longer

so

militarily

minded.[10] I

hope you

will

advance

well

in

your

career.

With

your downright

talent and

your

love

of

the

subject, I

should

think

so.

Accept my

warm

regards, yours,

Einstein.

In

any

case,

I

share

Weyl’s

conviction that it

must

be

possible

to

link

gravitation

and

electricity

to

a

unity;

I just

think that the

correct

link has

not

yet

been

found.[11]

566. From

Felix

Klein

Göttingen, 16

June

1918

Esteemed

Colleague,

I

discussed

your May

communication

now

variously

with

Hilbert

and

partic-

ularly

with

Runge[1]

but,

much

as

I

understand

your general

train of

thought, I

cannot

persuade

myself

of

its correctness. In

particular, I

do

not

see

how in

the

formula

Al

=

E

dx'4

dxß

dxn

dx,

Aa

Aß

you

can

replace

Aßa

with

Jß

.

Axa.[2]

So

for

the

time

being

probably nothing

will

come

of

a

lecture

by

me on

the

subject, especially

since in

the

coming

weeks

I