Yes, it is fair to infer that Mary and Sarah were twins.
The indexes were compiled quarterly, meaning that children born over nine months apart should appear in different quarters. If one was registered late then they often are entered in the index in the correct quarter as well, with a reference in the margin such as "see Dec '37".
The fact that both of them appear on the same page in the register suggests that they were registered at the same time. Births and deaths in England were originally entered in register books, very similar to those you might view when looking at baptism registers. When you order a "birth certificate" you are getting just one entry off that page extracted. Sometimes twins will appear on different pages in the register if, for example, the first twin is at the bottom of one page and the second twin is at the top of the next page.
The fact that Mary and Sarah were twins is further confirmed by the fact that both appear age 3 on the 1841 census.
It is worth checking the original baptism register as often twin births will be noted as such.
Often the time of birth is given in the birth register for twin births, mainly for inheritance purposes to know who was the eldest twin. It depends on the registrar whether this was done; clearly it was not in this case.
Keep in mind that with twin births it is not possible to infer the birth date of the other twin with absolute certainty. Usually twins were born on the same day, but I have a case in my tree where twins were born a couple days apart (the eldest twin did not survive).