Indeed the Peutinger Table, a medieval copy of a 4th or early 5th century map of the world, shows a "Temple to Augustus" at Muziris, one of the main ports for trade to the Roman Empire on the southwest coast of India. This and evidence of agreements for loans between agents, one of whom most likely lived in Muziris, and a rather oblique reference in the Periplus, all seem to point to a settlement of Roman subjects living in the region.
The most famous book about the presence of Roman colonists in India is "The Periplus Maris Erythraei: Text with Introduction, Translation, and Commentary" of Lionel Casson, published in 1989 (https://books.google.com/books?id=x5I4ZkR-VB4C&pg=PA11&source=gbs_toc_r&cad=3#v=onepage&q&f=false). In page 25 and 25 can be read the following:
At Muziris/Nelkynda there are unmistakable indications of a foreign colony. The clearest evidence comes from the Tabula Peutingeriana map. Nerxt to Muziris this map shows a building identified as Temp(lum) Augusti, "temple to Augusti"; such temple could only have been put up by Roman subjects living there
Almost as clear evidence is provided by the papyrus from the Vienna collection...it contains a reference to "loans agreements at Muziris", agreements between two merchants, one of whom very probably was resident at Muziris. And there it is an indication in the "Periplus" itself that points in that direction. The author states that Muziris/Nelkynda imported grain "in sufficient amounts for those involved with shipping, because merchants do not use it". The explanation can only be that the merchants "do not use it" because they are natives of the area and hence eat the local rice, whereas "those involved with shipping" are Westerners who prefer to eat what they have been accostumed to even though it means inporting it from thousands of miles away. These Westerners, permanently established, served as middlemen between their countrymen who arrived with the cargos and the local merchants
In the east coast at Arikamedu some two miles south of Pondicherry, archaeology has brought to light convincing signs of a colony of Westerners, an abundance of Roman pottery, especially Arretine ware, which reveals that its members were active since the first years of the first century on.....In addition, a passage in a Tamil poem attests to the presence of a colony of Westerners at a port on the mouth of the Kaveri river to the south of Arikamedu. It could well be that the Westerners resident in these places were chiefly engaged in forwarding goods not all the way to Egypt but only to associates stationed in Muziris/Nelkynda, who then sold them to the merchants from Roman Egypt.
There it is evidence for the presence in eastern India not only of Western merchants but of their wives as well...a statue found at Didarganj (a little north-east of Benares) portrays a young girl whose hairdo is strikingly similar to that on busts of Roman women of the Augustan period.
RELATED STUDIES & RESEARCHES
There are some studies about the possibility of the existence of one Roman colony (or more) in India. The best researches indicate that outside the extreme borders of their empire, the Romans created some small fortifications to defend the commerce of their merchants, mainly in the "Barbaricum" (as they called the areas populated by Germans & other tribes in Europe).
But also in the commercial route toward Asia the Romans made these small "castra": for example in the Farasan islands facing Yemen recently have been discovered the remnants of a legionaries cohort castrum, that was active in the first centuries after Augustus and that probably protected the area from pirates (https://books.google.com/books?id=gzJf1KgqrWQC&pg=PA80&lpg=PA80&dq=romans+in+farasan+islands&source=bl&ots=OkYC-Ssz-s&sig=ACfU3U07z3kMSckojzFPozE1H6VgACJ2DA&hl=en&sa=X&ved=2ahUKEwieruqaveTgAhWGylkKHbz0BK84ChDoATAGegQICBAB#v=onepage&q=romans%20in%20farasan%20islands&f=false. Some researches argue that something similar was probably created in the Socotra island, but nothing has been found until now; however it is possible: Roman control of the Farasan islands area was completely unknown until the 2004 publication of an inscription attesting to the presence of a detachment of the "Legio II Traiana Fortis" and auxiliary troops on the main island in 144 AD.
Furthermore Casson hinted that in order to defend the "Temple Augusti", located in the area of Muziris, probably some Roman legionaries lived there. Indeed southern India had "pirates", according to the famous Peutinger Table, and probably the Roman colonists -who Casson supposed had settled there- were protected by some kind of small fortification or castrum.
Indeed two are the India's city areas where it was more possible the existence of a Roman colony, between the first and fourth century during the heydays of the Roman empire: Muziris and Arikamedu.
MUZIRIS
"To those who are bound for India, Ocelis (on the Red Sea) is the best place for embarkation. If the wind, called Hippalus (south-west Monsoon), happens to be blowing it is possible to arrive in forty days at the nearest market in India, Muziris by name. This, however, is not a very desirable place for disembarkation, on account of the pirates which frequent its vicinity, where they occupy a place called Nitrias; nor, in fact, is it very rich in articles of merchandise. Besides, the road stead for shipping is a considerable distance from the shore, and the cargoes have to be conveyed in boats, either for loading or discharging. At the moment that I am writing these pages, the name of the King of this place is Celebothras". Pliny the Elder: The Natural History
Muziris was the "Primum Emporium Indiae" of the ancient authors like Pliny the Elder (who wrote just after Augustus times the above excerpt). It has been initially identified with modern Kodungalur or Cranganor (though the exact location of the harbor and city is not known, because of heavy alluvial deposition) in the Kerala State of southern India.
Later, a series of excavations were conducted at the village of Pattanam near Cochin by "Kerala Council for Historical Research" (an autonomous institution outsourced by Kerala State Department of Archaeology) in 2006-07: it was announced that the lost "port" of Muziris was found (https://www.thehindu.com/todays-paper/tp-national/Pattanam-richest-Indo-Roman-site-on-Indian-Ocean-rim/article16589683.ece), but some historians and archaeologists criticised this declaration. However, the last field report on the excavations (2013) explicitly shows Pattanam as Muziris (http://kchr.ac.in/images/img/pattanam_7th_season_handbook1.pdf).
Significant is the fact that abundant Roman coinage has come to light in Muziris area. Furthermore 80 % of these pieces of metal unearthed from the whole of India have been found in the Kongu country (Coimbatore plateau). Actually, more than two thousand Roman coins have been collected in this region, most of them ranging from the first century BC to the fourth century AD, principally from the reigns of Augustus and Tiberius.
A part of the raw glass imported to Muziris followed the same route and reached the urban centers of another area where large quantitites of raw glass have been found in excavations (particularly in Arikamedu). Ceramics such as terra sigillata found at Arikamedu, now datable to the early first century AD, have also been recovered at Kodumanal, Karur and Sulur in the Kaveri valley. Recently, sherds of rouletted ware were unearthed at Vellalur, near Coimbatore, also known for a Roman coin hoard. All these discoveries indicate a possible presence of Roman merchants (and their families) in the area of Muziris (in actual Kerala).
Indeed in Pattanam Mediterranean contact is represented by readily identifiable ceramic material, including a Dressel 2-4 amphora, as well as other finds of imported Roman amphorae and related fine wares. Such artifacts date from between the late first century B.C.E. to the fourth century C.E. Remnants of Roman glass bowls, fragments of painted glass objects, and glass pendants discovered at the site are also suggestive of personal belongings rather than merchandise: this additionally confirm the possible existence of a Roman colony in Muziris.
Finally we have to pinpoint repeating that evidence further supporting the supposition that Muziris may have had a Roman ‘merchant colony’ comes from the "Tabula Peutingeriana", a medieval map depicting the Roman world as it was in the first century C.E. This map shows a building marked as "Templum Augusti" (Temple of Augustus) at Muziris. Historians like Young argue that such a structure would have been built only by subjects of the Roman Empire, likely ones who either lived in Muziris or who spent a significant portion of their time there. Young further argues that the presence of foreign merchants is supported by the Periplus, in a passage that mentions “enough grain for those concerned with shipping, because the merchants do not use it.” These merchants who do not use grain are thought to be Indians who would instead have eaten rice, whereas ‘those concerned with shipping,’ are thought to have been resident foreigners.
ARIKAMEDU
Arikamedu, one of the emporia of Roman trade on India’s Coromandel Coast, has come to be regarded as essential for the study of overseas commerce with the Mediterranean world during the so-called “Indo-Roman” trade period – a term that has prompted some controversy.
Indisputable evidence for commerce with the Mediterranean exists in fragments of transport amphorae, cups and plates of terra sigillata, ceramic lamps and unguentaria, blue glazed faience and glass bowls found at the site. Two-thirds of the amphora fragments found at Arikamedu during the 1941-1950 excavations come from wine jars, suggesting that wine was a principal commodity sent to India from the Mediterranean. Many of the fragments originated in Roman Greece, from the island of Kos, though fragments of Knidian and Rhodian amphorae have been found as well. Fragments of Koan amphorae, originating in Campania, have led to the suggestion that wine from Greece was later supplanted at Arikamedu by an Italian production.
In addition to wine jars, fragments of Spanish jars for garum sauce and olive oil have been found as well. Sherds of terra sigillata, a slipped Roman ware, found in the 1989-92 excavations at Arikamedu and dated to the first quarter of the first century C.E., are thought to represent personal possessions, novelty items, or gifts. While typically considered part of the assemblage indicating resident foreigners, a sherd found with “megalithic” writing has caused speculation that some terra sigillata pieces were sold, bartered, or gifted to the local population.
Ancient Roman pottery found in Arikamedu, India
Arikamedu was "a trading station to which goods of Roman manufacture were imported during the first half of the 1st century AD", according to Peter Francis. For him, the buildings of the northern section of ancient Arikamedu were probably inhabitated by "non-Indians" (https://books.google.com/books?id=zzZBdGQN_TIC&pg=PA30#v=onepage&q&f=false).
