@arjun
according to this source there should be 3 tables (E1, R1, E2) and the 4th table would be R2. Reasoning for having 3 separate tables for 1:m relation below:
https://www.cs.uct.ac.za/mit_notes/database/htmls/chp06.html#optional-for-both-entities
1 R1 m
E1 E2
m R2 n
Please imagine the lines and boxes above
In 1:m relation, because it hasn’t been specified if participation of E2 in R1 is total, we have to ideally assume it to be optional.
And because it is optional, the foreign key of some tuples (if E1 and R2 is combined) will remain null. (check it out with making an example where relation is 1:m and participation is optional).
And generally, we don’t want null values in our tables, therefore we should make 3 tables E1, R1 and E2 instead of just 2 by merging R1 and E2 (sorry for not providing any standard text supporting this null value argument, but if you google you will surely find, and also i remember hearing this in the nptel lecture series of dbms of iit kharagpur Prof. Partha Pratim Chakrabarti)