Haradrim Warriors

Again, like my commentary on European cultures, please remember that my commentary on Africa here is a summary of thousands of years of history across a hugely diverse continent. Which is then applied the best I can to a fictional universe. So there is going to be some mental gymnastics that must be performed. So to is simplification and generalization.
So, the Aztecs are the only culture explicitly mentioned as an inspiration for their look - which was for the movies. The inspiration for them in the books was vague "African" societies. Since it is extremely impractical to generalize about all of Africa, I will focus on the Zulu. The Zulu military under Shaka is what I am most familiar with. Tolkien was born in South Africa after the reign of Shaka, so he most likely would have been familiar with the tactics the Zulu people used too. So I will be blending Aztec and Zulu military tactics and weapons.
Now, the Aztec appear to be wildly different than the African societies - namely the Zulu I will be drawing on. Namely, the Aztec were much more of an urban society and was much closer to a cohesive empire than the African societies inspiring the book versions of the Haradrim. Granted, Shaka did lead to some changes in Zulu culture that moved them closer to the Aztec society. But there are still major differences. So I will basically be using the Aztecs to fill in holes where I don't think the two contradict.
Southrons mostly live nomadic, tribal lives moving from oasis to oasis in a vast desert. In this context, tribal wars between the many mutually hostile tribes and raiding was common. I deal with that style of fighting in the Wild Men section. As I think that carries over here quite well, just imagine that the Southrons would do the same thing. There is incredibly little hard information on the Haradrim. But the facts here suggest that the troops they would produce outside the oilphaunts would be light cavalry. But whatever infantry that would exist would probably be skirmishers.
For the most part anyway. I say that because it seems unlikely that the Wild Men would have much in the way of horses, camel or cattle. In African tribal wars, access to grazing territory or water were important. As too were disputes over theft of these animals. They were highly prized - so much that they were the measure of wealth. So there was just as much incentive to conduct a raid to steal them as there was to violently defend against these raids.
When fighting breaks out, I would imagine that the most extreme flexible response approach would taken. If we look at the Aztec Flower War and traditional African war fighting methods, the typical battles were more violent rituals or festive skirmishes rather than what we today would thing of as war. The Aztecs and the Zulu under Shaka did to conduct full scale wars of conquest, which united many different local tribes together for a single purpose. These appear to be quite brutal when they did happen. When the smaller battles fail to resolve matters, the massacring of entire tribes and villages did occur. Likely less than what the more sensational reporting attests to, but it happened. So low level disputes would be militarized, but nominally so. Should things escalate, it will do so quickly to the point bordering on genocide. After all, tribal groups do not have the resources or skills for sustained campaigns or occupations. Resolution of a full scale war needs to defeat the enemy so completely that no hope of the enemy returning is possible. Similarly, the resources needed for the attacking group would be taken by survivors. When you have a territory persistently on the edge, those extra people would most likely be seen as an intolerable burden. So these extreme measures would be as predictable as they would be horrific.
Now, for similar reasons, I think mobilization would mean war. No group so desperate they would go to war for resources and massacre everyone to gain complete control can afford to mobilize all their forces without going to war. Now, it must be noted that the defeats of one side in the ritual combat was normally enough to force them to given the victor their demands. All that would be needed in a lot of cases is mere demonstration of superiority. The full scale brutality that I think the Haradrim would be willing to resort to would be when this fails. But some fighting would occur at some level.
The speed of mobilization of Southron tribes would probably be the most impressively fast of all the groups in Middle Earth. Imagine if Rohan was just the local eodred without the centeal authority and you start to get an idea. All the people in a tribe would be close at hand at all times. So would all their equipment and animals. There would be no vast expanses that troops need to cross in order to form up. Nor would there be and communication or logistical issues slowing things down. On the other hand, there is little time to respond. It is unlikely that there is little strategic depth at the tribal level. Once a threat materializes, it is either respond rapidly or risk everyone dying. Sometimes that was merely packing up and moving away from the attacking group. But that prep work needed to be done as soon as possible after the discovery of the threat.
I would suspect that this same logic would result in a first strike posture. A tribe would not have the most amazing economic or manpower reserves to fall back on. It is either total victory now or a high risk of total defeat. Which is part of the reason for many night attacks on villages and massacres. So if a threat appears, allowing it to attack first might prove fatal. It would then likely strike as soon as it can. This means a counter force posture. This is partly due to the need to make sure that the enemy in question has their offensive power defeated as soon as possible.
But also the "value" targets would remove highly valuable spoils of war from the equation without any real gain. Yes, destroying the enemy' economic base would prove useful in long wars. But the tribes in question might not last long enough for those benefits to materialize. Similarly, they might need to use the enemy resources that counter value attacks destroy to rebuilt afterwards. Offensive operations would be launched to seize economic assets from the other side, so attacks against them right off the bat would be illogical regardless of the cause of the war.
To me, the Haradrim would take a low intensity conflict position. Their tactics, unit size and resource base make them ideally suited for this style of war. Their enemies, suffering from similar low numbers and poverty too, would be perfect targets. This means cattle raids, the village night attacks and ambushes. We would expect them to seek out and kill any enemy group they find smaller than the attacking force while avoiding the larger ones. Think of the Haradrim of using the guerrilla war and defeat in detail models to achieve victory.
So here is the thing about Harad. Because of the tribal nature of the people living here, this is still a first strike seemingly total war posture. But from the perspective of Harad as a whole, this is much more akin to a no strike posture.

Composition

Manpower

As for exactly how big this force would be I have no idea exactly. The best I have is the 18,000 from the book that includes them and the Easterlings. 1,000 humans and 20 oilphaunts are listed for the movies. Other articles suggest the 18,000 number is just for Haradrim. Given that two separate Easterling ethnic groups are listed for the movie numbers, I will assume that one third or 6,000 are Haradrim. I would assume based on the comments I have seen that these are the total of all fighting age men, including the very old and the very young – falling probably from 14 to 60. I would use 3 members per household to get family sizes that are not on the battlefield – the children, elderly, women and disabled. This gives us a range of 3,000 to 18,000 total Haradrim.
This seems plausible to me. The lower population densities of nomadic cultures, the poverty of Harad would naturally put a lower cap on total population. Add in the constant warfare and slave raids and things do not look good for demographic growth. Though this is also partially due to geography. Harad is reported to be roughly the size of Africa. It is highly unlikely that all of Harad was emptied for this for no other reason than the time that would take. So these are likely the tribes from Near Harad. Which would be the ones who were most directly colonized by the Black Numenoreans. Near Harad is also the region where the camels and oilphaunts are from. Here would be the tribes ones most directly threatened by Gondor and able to benefit from its defeat.
In combat, the soldiers would be mostly infantry. Mounted troops were seen as the elite, similar to the European knights in that regard. Then, this force was broken down further with one spear man per 3 archers. Looking ahead a bit, we see that 90% of armies appear to on average to be infantry with the rest mounted. For an 3,000 to 18,000 person force, this means between 300 to 1,800 cavalry, 675 to 4,050 spear men and 2,025 to 12,150 archers.

Equipment

As far as armor, they used bamboo as the basis, working in animal hides, stone, bone and tusks to add further protection and to intimidate their enemies. Corslets of bronze or brass scales was used too. Though I would suspect that these would be rare. After all, there is little to no known metal available. So it would likely be reserved for the highest ranking people. The fact that most of their people would be on top of massive oilphaunts means that they would be exposed to danger less as well, meaning less need for the protection.
Their shields were yellow and black with spikes. This though we can expect to be common. Probably made primarily of bamboo or wicker and covered in leather. This seems logical - the Zulu Nguni shields were similar. The Nguni shields used by the Zulu were large and heavy - and like the Haradrim shields, made of hide. Generally, they were oval shaped and 5 feet tall. If one does not have much in the way of other armor, a more impressive shield is needed. These were also painted with colors to indicate regiment, status and rank.
The aztec shields were functionally the same. They were highly decorated. Copper and gold were used as decorations, even turtle shells and feathers hanging from the bottom were used. The front was heavily painted - also intending to function as heraldry. Their construction was a bit different though, but still using methods the Haradrim can reasonably be expected to have. Maize cane was used to form the basis. Cotton or maguey fibers were interwoven into this frame. Leather was used as padding and to provide extra arrow protection.
It was generally best suited for defending against the javelin and arrow fire coming from their enemies. This shield was also applied offensively. Zulu warriors were trained to use the left side of their shield to hook the right side of their enemy's. This exposed their ribs to the sort of thrust the Iklwa was designed for. Or they would just bash their opponent. Interestingly though, they could also be used to hide the soldier. In one battle, the soldiers were advised to hide behind their shields. The approaching enemies thought the waiting soldiers were actually grazing cattle, thus allowing the Zulu to get closer than they would have otherwise.
Dhlomo, Herbert Isaac Ernest. "The Shield in Tribal Life" (PDF). Retrieved 23 September 2016.
This applies primarily to the infantry though. The ones used by the oilphuants would be a bit different. I would imagine that they would probably be small to medium size. Again, they would probably not be seeing direct combat too much, so big impressive shields would probably be overkill. The Zulu umbumbuluzo shield was the same basic design - just 3.1 feet tall and held in one hand. Which seems adequate.
Generally, these shields were highly prized. They were held by the chieftain and only distributed when battle approached. Their painting on the front had the "coat of arms" for the tribe. Because of this, the shields had important cultural significance. Shaka apparently inflicted severe punishments on his soldiers who lost theirs. Special cattle were bred to create better hides for these shields.
The Sokoto Caliphate had cavalry with either mail or quilted armor. They also put wads of paper inside for extra padding. This was apparently effective against swords, but not as good against arrows. Leather body armor, shields and helmets were made from hippo and elephant. Since all of these seem close to what the Haradrim already use and all the raw materials outside the hippo hide are available to them, I suspect this armor would be replicated by the Haradrim.
The Mossi used leather to cover most of their horses. The horses had cotton and kapok fiber based armor in Sokoto formations. We can also expect that many layers of cotton tunics would be worn in combat, as was common in real life African forces. Also common was the use of magic spells for protection.
Aztec warriors used similar armor made of quilted cotton armor meant to primarily protect the torso. It was one to two fingers thick. A salt water brine was used to harden this. The result was hard enough to withstand obsidian blades. But it being padded armor meant it was effective against blunt force too. In all, this was quite effective armor. Spanish troops actually adopted it over their heavier plate armor.
Phillips, Charles (2006–2007). The Complete Illustrated History of the Aztecs & Maya: The definitive chronicle of the ancient peoples of Central America & Mexico - including the Aztec, Maya, Olmec, Mixtec, Toltec & Zapotec. p. 94. ISBN 1-84681-197-X. OCLC 642211652.
Hassig, Ross. Aztec Warfare: Imperial Expansion and Political Control. p. 88. ISBN 978-0806127736.

Now, cotton would need to be grown, not something I would expect to be in abundance in a nomadic society. Nor would the salt water needed be common in a desert. Which means I suspect that this armor would be found - just for the elites and the tribes allied with Umbar.
The Aztec helmets would also likely be a common if specialized item. It was made of hardwood and crafted to look like animals. Howler monkeys and big cats or birds were common. For the same reasons, I suspect that the jungle tribes and/or elites would be the main groups using these.
July, Robert Pre-Colonial Africa, Charles Scribner, 1975

Weaponry

Their melee weapon of choice is the scimitar. Though swords are mainly found in the wealthier tribes able to trade with Easterlings. This would likely be limited though. Far more likely than this would be the Aztec Mācuahuitl, which is essentially a cricket bat with obsidian blades around the outside. Axes were common too. So too was the dagger with a 7-9 inch obsidian blade called the Tecpatl.
Due to their lack of resources in Harad, Southrons would primarily rely on bamboo spears. This fits with the Nguni stick fighting style - which has the fighter use two long sticks. One is dedicated to offense and the other defense. A shield is the only armor used and only in the Zulu version. The spears used appear to be on average weaker in most societies prior to the rise of the Zulu. These tended to be used for both thrusting and throwing. Such a weapon/fighting style would be quite effective for skirmishers.
Though in pitched battles, a two handed spear would be used for a spear wall. They would place their shields upright in front of them. This would protect the archers placed behind them. The weapon used by the Zulu during Shaka's reign was the iklwa, a short spear with 9.8 inch broad, sword like tip meant for thrusting. Traditionally, they used the Assegai, a longer throwing spear. In the javelin, the tips were fire hardened wood or iron. In the smaller ritualized battles, several of these Assegai would be carried.
Apparently African war clubs were surprisingly advanced and useful for such a blunt weapon. They could be thrown with enough force to break bones on contact. These then had sharp teeth added to them. One group of Portuguese were wiped out by the fast and accurate throwing of these clubs. The Zulu Knobkerrie version was essentially a walking stick with a huge knob at the end. It could be thrown with lethal force or used in melee combat like a mace. The Aztecs used a similar weapon using copper, stone or wood balls. I suspect that some tribes would use these alternatives if the resources at hand allows. One club based weapon used by the Aztecs was 1.64 feet long with knobs on each side and a pointed tip. Thus, it was useful for thrusting and bashing. Their Quauholōlli was quite good at breaking bones.
Thornton, John Kelly Warfare in Atlantic Africa, 1500–1800, Routledge: 1999 ISBN 1857283937
Bows would be made of wood and antlers in the wealthier tribes. Due to these limitations, the bows used had a normal firing distance of only 100 yards. Stone, tusk and obsidian are used for arrowheads. Though those used by Sudanic mounted troops were apparently quite good - even getting compared favorably to guns. Yet it appears that it was more common for bows to be weaker, with 40 pound draw weights being normal.
I have little information on the Haradrim or tpyical african bows. But I have a little more on Aztec ones. Those tend to be 5 foot self bows with animal sinew string.
Typically, the Sundanese had 10-20 arrows in their quivers. 40-50 appear to be more common though. No feathers were added for flight stability. Which meant accuracy went down. Aztecs tended to use animal hide quivers with 20 arrow capacity. As some African societies used similar quivers, I would imagine this would be the standard size. Arrow heads used obsidian, flint, or bone points. Typically turkey or duck feathers were used for fletching. Now, fletching does greatly help the accuracy of arrows, I would imagine that this is what the Haradrim would use.
To compensate, two arrows could be shot at once and poison was often added to their tips. Specifically the Strophanthus genus was used for that poison, which stun opponents by stopping their hearts. The poison allowed a strike anywhere on the body to be more damaging. Which reduced the need for higher power and accuracy. One did not need to directly hit vital organs or penetrate as deep with poison. Any hit would be effective. The Aztecs also used poison darts - just derived from tree frogs.
Blow guns, slings and dart throwers were used by the Aztecs. The darts used for war were more like big arrows - a whopping 5.9 feet long. Copper, Bone and obsidian were used for the heads. Sling ammo was made with stone or clay with flint flecks. These could be quite deadly. Spanish soldiers could get wounded by the hail of stone even with their plate based armor. Now, the huge variety of environments within Harad and the decentralized tribal nature of the Southron culture means that their is likely to be no consistency in weapons. So the self bow I discuss above will likely be the most common based on history. But these other Aztec weapons would likely occur as well. Just mainly in the tribes from the jungles that have not been dominated by the
Lances were used in mounted combat by the Mossi. Often, these would be tipped with poison too. But typically, bows and spears were more common in those using camel mounts.

Vehicles

The Southrons have the major advantage in that they have oilphaunts, which only really exist in the Harad regions.. Essentially they are real world elephants, just 8 times as big. They have 4 full sized tusks and 2 smaller ones. These guys are between 50 and 100 feet tall and weigh enough to shake the ground as they walk.
In combat, they were charged straight into the enemy army. Their size means speed as well as terror – something complemented by their shrieking cry they let out as they charge. On top Southrons put bamboo frames that men would fire arrows or throw spears from. So yeah, not only are enemy troops being charged by a massive elephant, they are being shot in the process. No wonder they were effective. Should this fail to rout the enemy, the main force would follow behind with spears, bows and swords. Though the movie versions show them as exclusively on the backs of the oilphaunts.
When we look at the use of elephants in real life war, this is not too far off from what the Southrons do. They would be kept at the center of the line in order to prevent enemy charges or to take part in their own. These charges were more impressive than even the Rohirrim – their size, speed and appearance understandably made people flee before them. Even veteran Roman soldiers fled. Not to mention spear walls were ineffective against them. The brute force of the charge and the swaying tusks of elephants would at least knock back pretty much anyone.
However, they did have major weaknesses. The main one was that they could easily panic. Skirmishers became good and causing pain or killing the drivers. This would cause the elephants (or oilphaunts) to run amok and even charging back into their own ranks. Flames could scare them, as would spears thrown into their flanks. Also used was lighting pigs on fire and herding them towards the elephants. So, while useful as shock elements, I suspect the oilphaunts would rapidly lose their value as the surprise wears off.
The smell also caused horses to panic. which means I would imagine that the movie tactics would be the main ones used. The books suggest normal horses used in mounted combat. But this seems unlikely when the biggest asset in one's arsenal is a creature inherently scary to horses. So I would suspect that light infantry or camels would be the main backup for the Southron Forces.
Beyond the issues of the smell of the oilphaunts, there are major issues in using horses for the Southrons. First, there is the scarcity of resources in a desert environment useful for making fodder. And the heat would prove problematic. Africa, which provides the inspiration for Harad, was also home to many diseases that killed horses. But those were the larger, "traditional" states with more developed agriculture sectors and urban centers. Many African state imported horses, but tribal societies like the Haradrim would struggle to do this.
Given this and the fact that we know Harad has camels, I suspect that they would be the mounts used. They are much better suited for desert conditions. Plus, they too caused alarm and disorientation in horses. In one battle, them simply showing up had enough of an impact to allow an outnumbered king to win.
Jim Hicks, page 21 "The Persians", Time-Life Books, 1975

Structure

Great rivalries were promoted between members of Shaka's army. Apparently, this got so bad that different regiments had to be maneuvered into position separately otherwise they might fight with each other. Apparently, this also led to some defeats. Overly enthusiastic commanders and warriors would prematurely attack. Even when explicitly ordered to hold back. Rorke's Drift and at Kambula were Zulu defeats as a result.
The most basic unit for cavalry in Mossi society was between 10-15 men recruited from the elite classes. In the Kongo region, 100-125 soldier infantry formations were the norm. For the Zulu, the regiments could be anywhere between 400 and 4,000 depending on the context. Each was led by a InDuna responsible for maintaining discipline and local command in battle. Generally, it appears that they reported directly to a single person. There does not appear to be a single "supreme commander" of all Zulu field armies.
Drums, plus horns made from tusk, were common in African armies. We can assume this would be carried over to the Southrons too. This is essential. Like most armies, without things like radios, coordination would be difficult. African armies would stagger their assaults in pitched battles. This allowed tired or depleted ranks to fall back and be replaced with fresher units. But this required the ability to order specific groups at the right time to pull off.
Sometimes, the battlefield communications could be quite complex. Sometimes bells would be placed on horses. Flags were used too. But like most armies of the day, there was little consistent ability to coordinate once battle was joined. The Aztecs used boys around 12 to act as porters and messengers working with the priests marching with the armies. The signal for Aztec armies to attack were sent via drums and conch shell trumpets.
War was the main tool through which Aztec commoners were able to advance in society. As a person succeeded in battle, they could advance politically and economically. Economic growth was also driven by expansion, which means war in militaristic societies. Those that failed in military campaigns faced rebellion from previously conquered tribes. Nobles meanwhile would begin distrusting the leadership - leading to assassinations in some cases.

Tactics

As noted in the Wild Men section, women and children would watch from the sidelines of tribal battles. Often these would take place at arrianged places and times. Things escalated typically from taunts to single combat. Then to throwing of the javelins and tentative charges. Rare was sustained, mass charges and even rarer were attempts at extermination of the enemy. The defeated side would often concede territory, give up cattle or ransom some captives. But that appears to be the worst that normally occurs.
One area in which the Aztecs would influence us here is again the Flower War idea. This is discussed more in detail with the Easterlings. But they are important to mention here. They were primarily ritual battles at set times and places. The goal was to secure enough sacrifices to appease the gods, not outright military defeats or conquest. But they could defeat enemies through attrition or demonstrating overwhelming power/skill.
Like the Wild Men, raiding was the more common tactic used that more closely resembles our more traditional notions of war. The cattle enclosures would be burned. Often the cattle would be driven off and some captives taken. Often the defenders would just pack up and leave. If the attackers did not occupy the area, the defenders would be back in a day or two like nothing happened.
Morris, Donald R. (1998). The Washing of the Spears. Da Capo Press. ISBN 0-306-80866-8.
Tighter formations were used to combat mounted charges while looser ones were used to attempt flanking actions. There was also a recognized difference between light and heavy formations too. Light units in the Kongo region were more common, prized skill more and prioritized loose formations. The heavy ones as one could guess were the opposite. They preferred discipline as the main virtue and had heavier and more defensive weapons. Meanwhile, they were both less common and used larger, more dense formations.
Generally, mounted troops would be the main elite force operating in conjunction with a larger massed infantry formation. Mounted troops would rely on ranged and raiding actions most of the time. They would be the ones sent to capture people for sale as slaves for example. Generally, scouts and local informers would report to commanders where a good caravan or settlement was. They would move single file attempting to exploit the terrain to obtain surprise. When it came time to strike, they would surround the target and seize as many heads of cattle and slaves they could get away with, then retreat. In bigger attacks, light infantry were used as a vanguard. Though the cavalry would abandon the infantry if things turned out badly. Infantry were typically commoners, so they were "disposable".
However, the pitched battle use of cavalry in Africa appears to be similar to that of Europe. The infantry would be placed front and center. Cavalry would be on the wings with archers behind the spears. The archers would soften up the enemy prior to attack. One or two passes by the cavalry were made where they would throw javelins at their targets. Then they would charge with lances as the spear men advanced.
This was modified somewhat by the Fulani people. They tended to be more infantry based. In their organization, the spearmen were up front, with archers behind. The difference came in the general purpose reserves placed in the rear. This is where the club, sword and ax wielding soldiers could be located in pitched battle. They would charge in as the general shock element. Though sometimes, a more limited arrow volley was used. Sometimes the spear men up front were the main shock element. This is easily combined with the Kongo formations. When these attacked, the light formations would be placed in the center to draw attention from the enemy as the heavy wings moved forward to attack the flanks. Essentially, the diversity of tactics allows for flexible responses based on local conditions should the commanders be skilled.
R. A. Kea (1971). "Firearms and Warfare on the Gold and Slave Coasts from the Sixteenth to the Nineteenth Centuries". The Journal of African History. 12 (2): 185–213. doi:10.1017/S002185370001063X. JSTOR 180879.
July, Robert Pre-Colonial Africa, Charles Scribner, 1975
The Zulu kept with this basic idea. Their elite veterans were placed in the center fighting as heavy infantry. Their role was to engage the enemy formation through direct, frontal assaults. The goal was to simply pin the enemy in place. Once that happened, the light infantry would conduct flank attacks. Ideally, this would lead to encirclement and destruction of the enemy. Generally, the javelins were used similar to the Roman Pila - thrown at the enemy to break up the enemy formations and hurt morale. The Iklwa was used for the melee combat. A reserve was placed behind the veterans in case the enemy threatened to break out. The application of Zulu tactics under Shaka appear to be innovative in their brutality. In the first known case of his "buffalo tactics", the horns enveloped the enemy soldiers. And the women who came to cheer the soldiers on. All were killed. Merciless destruction or unconditional surrender appears to be the only options granted to the enemies of the Zulu.
Aztecs used surprisingly similar. Bows and slings were used first. Then the elite heavy infantry moved in for the main attack. They too used their 5.9 foot long "darts" as the opening volley of their assault. The less experienced or revered troops were held back. The reserves were used to try and outflank the enemy. But once the battle was entered, it devolved into a chaotic melee centered on individual hand to hand fighting. The totally untrained and inexperienced were the last reserves sent in. They were usually not committed until victory was assured, mainly so that they could exploit the rout to get more prisoners.
The Aztecs were also known for feigned retreats and ambushes. Also common - especially in flower wars and routs - was attempting to wound rather than kill. These aided in capturing the enemy soldier. Since the Aztec religion requires sacrifices and social advancement was based on captives, there were powerful incentives for this.

Training

The Zulu are reported to have marched 50 miles in a single day - barefoot over hot rocks. There is debate as to whether this was actually done or not in actual campaigns. Some claim that to be more of a myth while others confirm the idea. Those that are skeptical suggest that the Zulu armies march between 8.7 and 12 miles a day, sometimes with two days of rest. Yet fitness is an important trait for a nomadic warrior culture. So for the Haradrim, it is possible that the warriors are conditioned in training to be able to handle the extremes the Zulu are reported to cover, even if that actual distance is not actually expected in war. This would allow them to march the full 12 miles of the more conservative estimates and still be strong enough to fight - or handle longer distances in emergencies.
In Zulu society, trainees would challenge each other to stick fighting duels. Refusing was a dishonor. Also - apparently, those who could not keep up on the Zulu marches were killed on the spot.
Aztec training started at 15. The training was conducted by the veteran soldiers. The teachers would watch over the children. Training focused on using their weapons and how to take captives. However, this training was fairly basic for lower ranking children. Richer nobles would actually used gifts and food to induce these veterans to give special attention to their children. This training got better as the payments increased. The result is warriors were better trained if their parents were richer.
Laband, John (1997). The Rise and Fall of the Zulu Nation. Arms & Armour. ISBN 978-1854094216.
Morris, Donald R. (1994) [1965]. The Washing of the Spears: A History of the Rise of the Zulu Nation Under Shaka and Its Fall in the Zulu War of 1879 (New ed.). London: Pimlico. ISBN 978-0-7126-6105-8. OCLC 59939927. OL 7794339M.
Knight, Ian; McBride, Angus (1989). The Zulus. Bloomsbury USA. ISBN 978-0-85045-864-0.

Logistics

Logistical Support

Wars would be short. Probably just raids and pre arraigned pitched battles between small groups of warriors. Which would end the dispute in question. Or the war would involve night time attacks aimed at wiping out the entire enemy camp. Neither of these require much in the way of supplies. We can assume that the local war lords would not plan for more than a single day of battle. In real life medieval African war fighting, logistics proved to be an issue for both attackers and defenders. Defensive fortifications were merely to buy people time to escape.
Now, often there was no logistical support for armies. So when the archers ran out of arrows, they would withdraw from the field. Which could be quick. The firing two arrows at a time overcame poor accuracy with volume of fire. Which as one could guess burns through ammunition faster.

Auxilia

However, the Zulu armies under Shaka had children serve this role. Both boys and girls were recruited at 6 to become "apprentice warriors". They would be expected to carry things like cooking equipment and extra food. In Aztec armies, children still training to go to war handled tasks like cleaning and setting fires. In battle the Aztec children would serve as shield bearers for commanders.

Upkeep

Those still too young to fight still marched to war, typically as the servants. In Zulu armies, boys as young as 6 were actually used to handle the logistics. They carried a lot of the equipment, such as pots and beds. Generally though, the supplies came from the land. Cattle were driven with the army to provide food. Sometimes, local canoes and porters were forced to transport supplies. But this was rare and not enough to replace more modern supply lines.

Recruitment

We can assume that there would be no centralized recruitment organization. Manpower would be local, short term levies by the local tribes. Most likely this would mean all but the most extremely young and old men in a society. Which in turn means that there would probably be no long term fighting. Men would be mobilized when the crisis arose and sent home when it was over.
The age grade system common in African societies was manipulated by Shaka so that it allows a more consistent supply of manpower for his armies. In his system, everyone became a soldier at 20. This was not really the same as a traditional standing army in the European sense of the word. But it was closer to this than the traditional Zulu customs. Those were more akin to the feudal levies - short periods of service to the local chieftains.
The reforms of Shaka had some pretty important implications for Zulu society. First, when people turned 20, they were expected to join their respective regiments. They would build their own stockaded living area to which they would report when called up to fight. The recruitment also involved swearing an oath to the king - not local chief. Then, they were not allowed to marry without expressed permission of the king. This was typically around 40 and strictly enforced. Shaka for example apparently executed women and their lovers should they become pregnant prior to marriage.
These innovations led to two trends that reinforced each other. The first is a centralization of power. The recruitment was based on age - not geography or tribal affiliation. Which means that it became harder for local authorities to retain their power bases. Especially when the oath required loyalty to a higher position. By forging the traditional, decentralized bands into a larger body under the king's authority, Shaka was able to increase the size of his own power base as well. Essentially, this gave him significantly more power.
The other important element to this is it took men out of the local economy more. Which naturally hurts living standards. When one now has a larger standing army and a need to get resources to feed its society, there is a natural temptation to use said army. Some historians suggest that part of the reason for Shaka's expansionism was the need for more resources to pay for the army he created.
The warrior societies of the Aztecs were based on captives taken in battle. Boys did not become "real men" until they captured at least one enemy soldier. The most prestigious were those who captured at least 6 with "dozens" of other heroic deeds recognized.

Overall training Level

Semi-trained

Assumed Veterancy

Recruit