1. The use of the term cohort stretched back centuries before Marius when referencing the formations of the Roman allied Socii, who provided their heavy infantry similarly to how the Romans built their legions (age and wealth classes), but because each independent city state was not pulling from a much larger population as a unified Rome, they provided their contingents in units of cohorts, which were then built into Ala, or "Wings", separated in a consular army as Dexter Ala ("Right Wing) and Sinister Ala (Left Wing). Note, these were legion sized, organized and armed near identically to the Romans.

Some historical sources, specifically Livius, state that in the 2nd Punic War, while fighting in Spain, Scipio Africanus was forming the Roman legions into cohorts too. While some claim this is just an anachronistic use of the word by someone who was writing in the very late 1st Cent BC (about two centuries after the 2nd Punic War ended), there is some other literary and archaeological evidence to suppose that Roman legions were occasionally organized into cohorts as well.

Note, the maniples never disappeared. The pre-cohortal Roman legion was formed from two centuries in a Maniple, ten to fifteen maniples in a class line, and three class lines of heavy infantry, plus the skirmishers (rorarii/Velite), plus the citizen cavalry (Equites). After Roman legions were reorganized into cohorts, the formations were no longer class based (hastati, Principe, Triarii), and seemed instead to be formed on cohorts (Caesar's 4-3-3 for instance).

There is actually no historical indication that Romans forming into cohorts had anything to do with Marius. Sallust, writing about the Jugurthine War, suddenly mentions mid battle that four cohorts conducted a counterattack during the battle of Muthul River, which occurred in 109 BC; while Marius was in attendance at that battle as a legate, the command was held by Q. Caecilius Metellus. Later, Sallust continued to use the term cohorts extensively to describe not only the Socii but also the Romans, which is why most historians date the time period of the cohortal structure to the late 2nd Cent BC.

Other modern historians say otherwise. For example, Dr. Michael J. Taylor believes the cohortal structure was a slow gradual process throughout the 3rd to 2nd Century BC, with some Roman generals using it, and some not, used primarily because it gave another tactical level to the Roman formation, larger than Maniple, smaller than Legion, for missions such as raiding, foraging, etc. Taylor also believes it would have also assisted in more quickly forming a triplex acies battle line from the army's march in column. Lastly, it almost certainly became standardized after the Socii War, when Rome no longer levied their Legions inside Rome itself, but traveled to various areas to recruit cohorts (similar to how they had previously organized the Socii), and then bringing the various newly raised cohorts together to then form them into legions. So recruitment likely played a major part of it, after Rome enfrachised all free Latin and Italic people of the peninsula south of the Po River.

2. There is no historic evidence of there being any tactical subunit under the century level. We do know that Romans had mess group called the Contubernium, the "tent section," which was roughly 6-10 ten soldiers and a servant. Marius is stated to have streamlined the baggage train, removing extra servants and mules that previously had belonged to any individual soldier who brought them, forcing a tent section of common soldiers (this did not apply to any officers) to share one mule and one servant for carrying the tent section baggage. However, there is no evidence at all, even secondary, that during the Republic or Principate era that they served any tactical function.

3. At the battle of Ilerda during in Spain campaign during the Civil War, in his commentaries, Caesar described a triplex acies formation formed as each legion having four cohorts in the first line and three in the second and third line. The reason it couldn't be even, as it had been during the "Manipular" time period, was ten cohorts is not divisible by three lines, they cannot be even. With a duplex acies, ten cohorts is divisible by two two, so five cohorts in the front and rear is possible, probably likely. But there are no sources that describe it in detail. Realistically, it would be up to the general in how he ordered them being formed, it would depend on the conditions, the terrain, the enemy, the frontage he needed to hold, and whether or not he needed a second line of equal size to reinforce the front, or just enough to augment some of the front line, not relieve them all.

4. There were no squad leaders in the Republic and Principate. Vegetius, writing much later, describes Decanus, but either he got things wrong or he was writing of a different age. That said, if the Romans did have a privileged rank under the principales positions (Optio, tesserarius, signafier), they would likely have served in the front rank closer to the enemy, led a "file" (if they even existed, which there is no proof of during the Republic or Principate), or served in the right of their respective ranks, as the right was the position of honor.

5. It really depended on what role they served. Auxiliary cavalry generally fought in a manner that was culturally significant, Rome did not force them to change, not during the Republic at least. A Gallic cavalryman, would be using a four horned saddle, clad in mail, a stout helmet, wore spurs, carried an oblong or hexagonal center grip shield with his left hand or strapped over his shoulder, had a sheathed long sword, a fighting spear for poking, and maybe a lighter throwing spear or javelin. Meanwhile, a Numidian cavalry wore almost almost no armor, did not use any sort of saddle or even any tack except a simple rope collar and a stick. An Armenian or Asian Greek cavalryman might be kitted out as a cataphract. It really depended on the nationality of the auxiliary unit.

6. The Greek historian Polybius described Roman soldiers carrying two pila, a heavy tanged version and a lighter socketed version. While marching, they likely carried them strapped to their furca, the long pole they held their baggage on that rested on their shoulder (like a tramp's stick), and in battle they would have held in their throwing hand while walking. While throwing, they could either transfer one of them to the shield hand, using the thumb to temporarily pinch it while throwing the other, or stick it in the ground with the butt spike, toss one, then grab the other. However, other sources don't mention multiple pila, and many iconography don't show two, only one, including monuments, funerary stele, etc. So they might have only carried one. Their tactics called for at least one, by the Late Republic, nearly all charges were done by a massed pila volley and then drawing swords to charge into the disorder the pila volley caused.

For more info, I recommend the Osprey book by Ross Cowan, titled Roman Battle Tactics: 109 BC - AD 313, which accurately describes this very period (the so-called Marian Reform time period).