DOCS.

486,

487

MARCH

1918 503

professors

is

dated

20

Nov[em]b[e]r

1913

and bears the

file

number

U. I. K. No.

4054.[8]

With that

you

will

easily

steer Mr.

Kruss

on

to

the

right

track. If in

the

course

of

the

negotiations you

should believe

that

you may

in

any way

be in need

of

my involvement,

I

naturally

am

always

at

your disposal.

With

cordial

regards,

yours,

Planck.

487. From

Felix Klein

Göttingen,

20

March

1918

Esteemed

Colleague,

Here

is

my

long

reply

now

to

your

v[alued]

letter

of

the

13th.[1]

First

of

all,

my agreement

with

your closing

remark:

I

had noticed at

the

t[ime],

during

proofing,

that

my

statement

on

p.

13 of

my

note[2]

that the

special

theory

has

10 field

equations

Kuv

=

0

could

seem

misleading

(because

there

are

the

20

equations

(uv,

pa)

=

0,

of

course),

but

had left it because

the

situation is

described

clearly

at

the bottom

of

p. 5.[3]

Then:

on page

9

of

my

note,

l[ines]

3,

4

from

above,

the

/g

(which was

deleted

in

the

separatum

sent

to

you)

must be

retained

after

all[4] and,

as

I

now see, a

minus

sign

must be inserted before

the

summation

signs.-

In

other

respects,

however,

I want to

stand

by

the considerations

of No. 9

of

my

note[5]

and

substantiate

them here in

that

I

compare

them

formula

by

formula

to

your paper

on

Hamilton’s

principle, etc.,

in

the Berliner

Sitzungsberichte

of

1916:[6]

1)

My

K's

and

aQ's,

multiplied by

/g,

are

directly your

G’s

and

M’s,

insofar

as

I

disregard

the

special

form

of

Q,[7]

to which

I

restrict

myself

in

my

note

but

which

are

used

only

in

the

definition

of

6qp

in

f[ormula]

(13)[8]

and in

the

transition

from

(14)

to

(14').

2)

My Lagrangian

derivatives,

Kuv,

aQuv,

again

multiplied by

/g,

read for

you

as:

dG*

_

_d_

f

dG*

\

dM

dg^u

dxa \dgliu) ’

dg^v

For

the

Lagrangian derivatives,

there

is

indeed

no

difference in

operating

with

G

or

with

G*.

3)

Instead

of

Kuv,

aQuv,

I

can now,

in order

to

come

closer to

your designation,

introduce

“mixed”

components

as

needed:

K

=

j9-Y,Kv,gr

aQ:

=

a^g-J2Q^.