# The Scattering of Cosmic Rays by the Stars of a Galaxy

> An annotated edition of this paper on Fermat's Library, where scientists and readers explain its ideas, equations and history in the margins. Annotated edition: https://fermatslibrary.com/s/the-scattering-of-cosmic-rays-by-the-stars-of-a-galaxy

- **Authors:** Richard P. Feynman and Manuel S. Vallarta
- **Published:** 1939
- **Topic:** Physics
- **Original:** https://s3-us-west-1.amazonaws.com/fermatslibrary/uploads%2F1479615750891-vhlzgimmn2d-299c61f9cd9a899ec45adaa2a0e5a051%2Ffeynman_vallarta_first_paper.pdf
- **Pages:** 2
- **Annotations:** 4 comments in 4 threads
- **Featured:** Fermat's Library Journal Club
- **Annotated edition:** https://fermatslibrary.com/s/the-scattering-of-cosmic-rays-by-the-stars-of-a-galaxy

## Annotations

### Annotation 1 (page 1)

Beside the passage: “The Scattering of Cosmic Rays by the Stars of a Galaxy”

**Annotation** by Micael Oliveira (2016-11-20):

> **This paper is Feynman's first published work.**
>
> The first official scientific publications by Richard Feynman first appeared in 1939 in the Physical Review. He published two papers as an undergraduate student:
> - "The Scattering of Cosmic Rays by the Stars of a Galaxy", co-written with Manuel Vallarta
> - ["The Forces in Molecules"](http://faculty.washington.edu/miller/51816/Feynman.pdf), Feynman's senior thesis
>
> This paper was published in 1939 in his final year as an undergraduate at the Massachusetts Institute of Technology (MIT).
>
> Feynman and his Professor Manuel Vallarta were investigating whether cosmic rays came from inside or outside the galaxy by studying their scattering by the magnetic fields of the galaxy's stars.
>
> Their result is simple and straightforward but their reasoning is very clever: they stated that the probability of a particle’s emerging from a clump of scattering matter in a certain direction must be equivalent to the probability of an antiparticle's taking the reverse path. (like an anti-particle going backwards in time)
>
> Feynman went on to receive his Bachelor's degree from MIT and attained a perfect score on the graduate school entrance exams to Princeton University in physics.
>
> !["Feynman 1939"](https://qph.ec.quoracdn.net/main-qimg-5f3903392fe56957ff50821bff2186ed-c?convert_to_webp=true)
> *1939 - Picture of Richard Feynman at the Princeton Library*

### Annotation 2 (page 2)

Beside the passage: “any extended distribution of matter, that is, F(R, P; R', P') = F(R', P'; R, P), where F(R, P; R', P') is the probability that a particle going in the direction R at the point P will emerge in the direction R' at the point P'. This is because the probability of following any route is equal to the probability of …”

**Annotation** by Hemant Singh (2016-12-07):

> This paragraph displays genius of Feynman. No equations or laws, just defining the problem in a better way, with making reasonable assumptions or ruling out possibilities, and he reached an equation. Pure delight!

### Annotation 3 (page 2)

Beside the passage: “any extended distribution of matter, that is, F(R, P; R', P') = F(R', P'; R, P), where F(R, P; R', P') is the probability that a particle going in the direction R at the point P will emerge in the direction R' at the point P'. This is because the probability of following any route is equal to the probability of …”

**Annotation** by Micael Oliveira (2016-11-20):

> !["Galaxy with N stars"](http://i.imgur.com/ZUAzqqe.png)
> **Figure 1:** *The dashed line represents an element of Volume $dV$, the black dots represents the stars. The solid line depicts the trajectory of a cosmic ray inside that element of volume when it is scattered by a star.*
>
> Consider a set of stars (dipoles) distributed inside an element of volume $dV$.
>
> We assume that all the stars are:
> - non-absorbing: we consider the CRs have sufficient energy to collide with the star to be negligeable
> - non-capturing: the set of these type of orbits is negliegeable compared to all the other orbits
>
> Thanks to these assumption we know that all particles entering the distribution of matter will emerge at infinity, thus we have summing over all the possible directions $R'$ we can write:
> \begin{eqnarray*}
> \int F(R,P;R',\infty) \, dR' = 1
> \end{eqnarray*}
>
> We want to find the intensity of particles in the direction $R$ at point $P$: $I_p(R)$. Considering a beam of particles at infinity with constant intensity
> $$I_{\infty}(R') = I_{\infty}$$
> (independent of R') and taking into account that the probability of following any route is equal to the probability of following the reverse route we can write that:
>
> \begin{eqnarray*}
> I_p(R) &=& \int F(R',\infty;R,P)I_{\infty}(R') \, dR' \\
> &=& I_{\infty} \int F(R,P;R',\infty) \, dR' \\
> &=& I_{\infty}
> \end{eqnarray*}
>
> We can conclude that the intensity in any direction at any point $P$ is the same and the distribution is isotropic at $P$ if it is isotropic at infinity.

### Annotation 4 (page 2)

Beside the passage: “From the remark made previously, it is clear that if the distribution of positive and negative particles at infinity is isotropic, it will also be isotropic at any point P, except for small irregularities due to absorption by collision and by capture into periodic orbits. We conclude that particle scattering by …”

**Annotation** by Micael Oliveira (2016-11-20):

> Feynman and Vallarta found that the **net effect of the scattering of cosmic rays by stars was zero.**
>
> They have shown that  that the intensity of cosmic rays is isotropic and thus the magnetic fields of the stars are unable to contribute anything to the solution of the problem whether or not cosmic rays come from beyond our galaxy.
>
> If cosmic rays seem to come from all directions, it is not because the stars are interfering and disguising their original orientation.

## Full text

Extracted from the PDF. [Annotation N] marks where each annotation above sits.

#### Page 1

LETTERS TO THE EDITOR

Lifetime of the Yukawa Particle

Recent investigations by various authors' have made it very probable that the hard rays of the cosmic radiation (mesotrons), now identified with the particle of Yukawa' of mass p~200m (m=mass of the electron), are unstable and will decay spontaneously into electrons and neutrinos. The lifetime for a mesotron at rest has been estimated from experience to be of the order 2-4&(10 ' sec. Yukawa himself calculated the lifetime on the basis of his ideas to be of the order 0. 25&(10 ' sec. , a result not far from the observed value. However, the present author' obtained on the same assumptions a much smaller value. The importance of this question may justify a restatement of the theoretical result and an explanation of this difference.

The final formulae for the lifetime obtained by both authors is the same, apart from differences in notation. It can be written in the form

G2 m 4h r — — 4~~ Ac p pcs Gy~

In this formula k, m, c have the usual meaning, and p, is the rest mass of the mesotron. G is the constant of dimension of a charge in the potential between nuclear particles V(r) = (G'/r)e rw&l& following from Yukawa's theory. G'/kc is of the order' 4 p, /3II (&=mass of the proton) but probably somewhat larger than this quotient. The lifetime r is therefore essentially proportional to p 4. Gz finally is the constant in Fei:mi's theory of P-decay, normalized to be a pure number. The form of interaction assumed for the coupling between proton, neutron and the electron neutrino field is

G~mc'(&/mc)'(4N*PPI ) (v.*Ps. )+c. c.

(f~, P~, p„, q, being the wave functions of neutron, proton, neutrino and electron, respectively). This leads to the probability for emission of an electron of energy e

Gy' mc' (ep — e) (e mc ).&de w(e)de= ( M['

(2 )' a (mc2) 5

where ep is the maximum energy of the emitted electrons and M a matrix element from the motion of the heavy particles inside the nucleus. The discrepancy in the calculated lifetimes comes from the different values used for the constant G~. As discussed by Bethe and Bacher' and by Nordheim and Yost, ' the experimental value of Gz depends quite appreciably on the group of elements which are taken for comparison, the difference being due in all probability to the matrix element M; which is smaller than unity for heavy elements but can be expected to be unity for light positron emitters. The value for Gz used by Yukawa (0. 87)&10 " in our units) corresponds to the heavy natural radioactive elements, while the value deduced for the light positron emitters' is GJ =5. 5&&10 ". It seems beyond doubt that this later value has to be taken for our purpose. With the present most probable values G'/Ac = 0. 3; p=200m; Gg=5. 5)&10 ', we obtain from (1) v =1. 6&(10 ' sec. , i.e. , a value about 10 ' times too small. A decrease in

the assumed value for p, to 150m would increase v only by a factor of order 3. In view of this definite discrepancy the question arises whether any modifications of the theory could give a better result. It'is to be noted firstly that the introduction of the Konopinski-Uhlenbeck form of the P-decay theory would only make matters worse as it would introduce roughly another factor (m/p)'. A real improvement can only be expected by a complete reformulation of the theory. One possible suggestion would be to assume that the disintegration of a free mesotron is in first order approximation a forbidden transition, while in nuclei it is made allowed by the influence of the other nuclear particles.

L. W. NoRDHEIM Duke University, Durham, North Carolina, February 14, 1939. ~ H. Euler and W. Heisenberg, Ergebn. d. Exakt. Naturwiss. (1938); P. Blackett, Phys. Rev. 54, 973 (1938); P. Ehrenfest and A. Freon, J. d. Phys. 9, 529 (1938); T. H. Johnson and M. A. Pomerantz, Phys. Rev. 55, 105 (1939). 2 H. Yukawa and others, I — IV, Proc. Phys. Math. Soc. Japan 17, 58 (1935); 19, 1084 (1937); 20, 319, 720 (1938). o L. W. Nordheim and G. Nordheim, Phys. Rev. 54, 254 (1938). 4 R. Sachs and M. Goeppert-Mayer, Phys. Rev. 53, 991 (1938). o H. Bethe and R. Bacher, Rev. Mod. Phys. 8, 82 (1936). 6 L. W. Nordheim and F. Yost, Phys. Rev. 51, 942 (1937). It has to be noted that the formula for 7 o on p. 943 should be ro ~ = (Gg2/(2x)3) )&(mc~/$). The value of Gg is then determined from the empirical value vo & — 10 4.

The Scattering of Cosmic Rays by the [Annotation 1] Stars of a Galaxy

The problem dealt with in this note may be formulated in the following way: imagine a galaxy of N stars, each carrying a magnetic dipole of moment p, „(n = 1, 2, . . . N) and assume that the density, defined as the number of stars per unit volume, varies according to any given law, while the dipoles are oriented at random because of their very weak coupling. Under this condition the resultant field of the whole galaxy almost vanishes. Let there be an isotropic distribution of charged cosmic particles entering. the galaxy from outside. Our problem is to find the intensity distribution in all directions around a point within the galaxy. Its importance arises from the fact that if the distribution should prove to be anisotropic a means would be available for determining whether cosmic rays come from beyond the galaxy, independent of the galactic rotation effect already considered by Compton and Getting. ' Suppose we consider a particle sent into an element of volume d V of scattering matter in a direction given by the vector R. Let the probability of emerging in the direction R' be given by a scattering function f(R, R') per unit solid angle. Conversely a particle entering in the direction R' will have a probability f(R', R) of emerging in the direction R. Let us assume that the scatterer (magnetic field of the star) has the reciprocal property so that f(R, R') =f(R', R). In our case this property is satisfie provided the particle's sign is reversed at the same time as its direction of motion. That is, the probability of electron's going by any route is equal to the probability of positrons going by the reverse route, If it has the reciprocal property for each element of volume it will also have it for

#### Page 2

LETTERS TO THE EDITOR

any extended distribution of matter, that is, F(R, P; R', P') = F(R', P'; R, P), where F(R, P; R', P') is the probability that a particle going in the direction R at the point P will emerge in the direction R' at the point P'. This is because the probability of following any route is equal to the probability of following the reverse route, through the same elements of volume. Thus the probability of a certain end result from a number of possible routes will equal the probability of the reversal of the result occurring through the reverse [Annotation 2] routes, In our case the scatterer (star) is to a large extent nonabsorbing and noncapturing. The former is true except for particles colliding with the star, which can only happen when their energy is sufficiently great, and the latter is true except for particles which follow asymptotic or -periodic orbits in the magnetic field of any one of the X stars. These orbits, however, almost certainly form a set of zero measure in the manifold of all possible orbits, ' that is, they occur only exceptionally. Thus, while a dipole magnetic field can imprison charged particles starting from a point within it and can also keep them away if starting from infinity, depending on their energy and angular momentum, it can only exceptionally capture such particles starting from infinity. In our case, therefore, all particles starting in a direction R at a point P, sufficiently far from all neighboring stars, have only a small chance of being either absorbed or [Annotation 3] captured in a periodic orbit (of finite or infinite period), so that the great majority of them will emerge at infinity. For almost all particles, therefore, the probability of emerging at infinity must be unity, or

J, F(R, P; R', ~ )dR' = 1

almost always. Now consider a beam of particles at infinity whose intensity in a direction R' is I (R'). The intensity at P observed in the direction R will be

I„(R) =

f F(R', ~; R, P)I„(R')dR' (2).

Using (1) and assuming an isotropic distribution at infinity such that 1„(R') is a constant (independent of R'), we find that Eq. (2) becomes

Ip(R) =I J F(R, P; R', ~)dR'=I„(3)

by (1). Therefore the intensity in any direction at P is the same and the distribution is isotropic at P if it is isotropic at infinity.

From the remark made previously, it is clear that if the distribution of positive and negative particles at infinity is isotropic, it will also be isotropic at any point P, except for [Annotation 4] small irregularities due to absorption by collision and by capture into periodic orbits. We conclude that particle scattering by magnetic fields of the stars is unable to contribute anything to the solution of the problem whether or not cosmic particles come from beyond our galaxy. The considerations developed in this note clearly hold so long as the scattering centers satisfy the conditions of being nonabsorbing and noncapturing, irrespective of the law of force which is responsible for the scattering. It. need hardly

be emphasized that they apply only to the case in which there is no resultant magnetic field for the whole galaxy, such as would exist if the dipoles were oriented along preferential directions. In this case particles would either be imprisoned if born within the galaxy, or kept out, if coming from outside, depending on their energy and angular momentum. The reciprocal property of paths would then break down in general, but would still hold for any allowed direction at any point within the galaxy. M. S. VALLARTA R. P. FEYNMAN Massachusetts Institute of Technology, Cambridge, Massachusetts,

February 15, 1939. ' A. H, Compton and I. A. . Getting, Phys. Rev. 47, 817 (1935). M. S, Vallarta, C. Graef and S. Kusaka, Phys. Rev. 55, 1 (1939). ~ See the discussion by E. J. Schremp, Phys. Rev. 54, 153 (1938); and forthcoming papers by O. Godart and by A. Banhs, Jr.

Nuclear Excitation of Indium by X-Rays

It has been shown recently'' that the stable nucleus In"5, when excited by fast neutrons or protons, may be left in a metastable excited state, designated by In"'*, from which it decays, emitting negative electrons, with a half-life time of 4.1 hours. We have now observed that the same metastable state can be excited when indium is irradiated by x-rays. The x-rays were produced by bombarding a 2-mm thick lead target with electrons from an electrostatic generator. A thick indium foil, 1 inch in diameter, was placed directly behind the lead target. After 30 minutes irradiation at an electron energy of 1.73 Mev and a current of 10 ya, the indium foil showed an initial activity, recorded on a Geiger-Muller counter, of 45 counts per minute. The activity. decayed with a period of approximately 4 hours. The walls of the counter reduced the intensity of the rather soft P-rays to about one-half. Until more is known about the effective x-rays, no well-defined cross section can be deduced from these data. By varying the bombarding voltage it was established that the eAect has a threshold at 1. 35&0. 1 Mev. This result might be interpreted by assuming that In"5 has at that energy an excited state which combines both with the ground state and the metastable excited state. In a note which has just become known to us Pontecorvo and Lazar' also report the excitation of indium by x-rays. In their experiments the x-rays were produced with an impulse generator working at a peak voltage of 1850 kv. GEQRGE B. CoLLINs BERNARD WALDMAN EDWARD M. STUBBLEFIELD Department of Physics, University of Notre Dame, Notre Dame, Indiana.

1VI. GQLDHABER Department of Physics, University of Illinois Urbana, Illinois, February 15, 1939. ~ Goldhaber, Hill and Szilard, Phys. Rev. 55, 47 (1939). ~ Barnes and Aradine, Phys. Rev. 55, 50 (1939). ~ Pontecorvo and Lazar, Compte rendus 208, 99 (1939).

---

Source: [Fermat's Library](https://fermatslibrary.com/) — https://fermatslibrary.com/s/the-scattering-of-cosmic-rays-by-the-stars-of-a-galaxy. More annotated papers: https://fermatslibrary.com/llms.txt
