The simple answer is that no-one alive today knows. In fact, it’s likely no-one has ever known, even at the time. There were censuses taken, but they did not consider religious affiliation and so any attempt to answer this question (or at least the first part), is going to be pure guesswork. All is not lost however, because we can have a go at guessing and some guesses are going to be more reasonable than others. The evidence at hand would include references made in texts, archaeological findings, and such records as do exist.

Let’s start with a word on censuses and general demographics. It is impossible to speak with any confidence about population figures for the Roman Empire. The census data for the late Republic are considered reliable and yet even here population estimates range from four million to up to 10 million in Italy. The issues include how to interpret the data - do the numbers in the censuses refer to everyone, to households, or only male citizens? When it comes to the empire at large estimates range from (at the time of Augustus) 45 million up to 130 million. This is, as one scholar puts it, differences between historians that are not easily resolved being ‘orders of magnitude’ apart. Furthermore, going from the first to the third century can see historians predicting either increases or decreases in the population. For many scholars, the uncertainty in this topic means that they just steer clear of it altogether.

Next, to Christianity. Given what happened with Christianity in the end, it would not be surprising to find some interest in this question. However, given what I said above, not many scholars have dedicated time to it and I suspect the reason is that it is considered so speculative that there is not much value in it. The guru for the numbers of Christians is Rodney Stark. In the late 1980s and into the 1990s he pioneered sociological approaches to the growth of Christianity. Stark projected an initial 1000 Christians at the death of Christ and worked out that if there was 40 per cent growth per decade, we would get to 10 per cent of the empire’s population by around AD 300. This argument was supported by the fact that these numbers match the growth figures of new religious movements in the modern period, such as Mormonism, giving them an air of plausibility. Pretty much every historical reference today regarding the numbers of Christians in the Roman Empire quotes from Stark and to my knowledge no-one else has looked into the topic in any real depth since.

If we now take the typical average estimate for the population of the Roman Empire, 60 million, and the 10 per cent figure, then there were about six million Christians at the time of Constantine’s legalisation and subsequent adoption of Christianity. Estimates for the numbers of Jews are broadly similar, two million to 10 million, or three to 17 per cent. If we go for a number in the middle and get 10 per cent, like the Christians, then a whole fifth of the empire follows one of Abrahamic faiths. The other 80 per cent would mostly be polytheistic and include local religions alongside Roman paganism. There were other religions around at the time as well as these, Mithraism for example, which was popular in the Roman army.

Before we finish, there are a couple of case studies that seem to suggest that Stark’s percentages hold up pretty well even if our confidence in the raw numbers is low. They also throw some light on the other religions. The first is the town of Dura-Europos, the second the comments of Bishop Cornelius. Both are dateable events to the 250s.

Dura-Europos was a garrison town abandoned in 256/257. It had a population of about 5000 people and was probably typical of many towns in the eastern part of the empire, albeit more affiliated to the military. It contains at least 14 religious buildings, including a christian house church (domus ecclesia), a synagogue, a mithraeum, and temples to Roman, Greek, and ‘Eastern’ gods. Both the church and synagogue were converted private houses. The church could hold around 75 people at any one time, while the synagogue was bigger with seating for around 120. If we assume these figures to represent the full populations of each respective religion, the proportion of Christians is 1.5 per cent of the town and of Jews 2.4 per cent. Stark’s estimates of Christians for 250 are 2 per cent of the population.

Meanwhile, Bishop Cornelius of Rome gives us some details about the church there in some of his letters. He tells us the number of deacons, dependents, and other clergy and analysis of this has resulted in an estimated Christian population of 30,000-50,000. The city of Rome was declining from a peak of around one million in the first century AD, and leaves us in the five per cent range.

These two examples actually fit quite well with our understanding of christian demographics. It was predominantly an urban religion, well established in the major cities, but the majority of the empire's population was rural (even though it was highly urbanised compared to the early medieval period). It is impossible to say how representative these examples are, but if one was to extrapolate five per cent to the cities and 1.5 to the countryside, you would end up with around three per cent of the overall population in the 250s being Christian. Consequently, the growth to 10 per cent by 300 is plausible. For Judaism, the numbers were greater in the east and so a larger population in Dura-Europos would be expected and is suggested by the evidence.

There’s more that could be said here, and many other questions. The reign of Constantine is particularly interesting because after he converts all subsequent emperors bar one are Christian and most assume that the growth in Christian numbers were exponential in the fourth century with around 90 per cent of the population Christian by the year 400. It is hard to know whether he adopts Christianity because of the exponential growth that is assumed, if the exponential growth results from his conversion of the empire, or if the growth is exponential at all.

I will have a go at the second part of the question in another post and we’ll return to Dura-Europos to help us get an answer.