The middle film of the trilogy. The Battle of Helm's Deep, Gollum's introduction, and the structural sequel against which all subsequent ambitious sequels are still measured.
Following the events of The Fellowship of the Ring, the Fellowship has fragmented. Frodo and Sam continue toward Mordor, joined unwillingly by Gollum/Sméagol — the creature whose previous ownership of the Ring has destroyed him across centuries. Aragorn, Legolas, and Gimli pursue the Uruk-hai who have captured Merry and Pippin. The captured hobbits escape into Fangorn Forest, where they meet Treebeard and the Ents.
The film tracks three parallel narrative threads across roughly six days. Aragorn's group arrives at Rohan, where King Théoden has been corrupted by the agent of Saruman. Théoden is restored and leads his people to the fortress of Helm's Deep. The film's third act is the night-long Battle of Helm's Deep — 10,000 Uruk-hai against a few hundred Rohirrim, with Gandalf's cavalry arrival at dawn breaking the siege. Frodo and Sam's story closes on the cliffhanger of Frodo deciding to trust Gollum despite Sam's warnings.
The Two Towers had an unusual production challenge: it is, by structural necessity, a middle film. It has no real beginning (the audience is dropped into action continuing from Fellowship) and no real ending (the story continues into Return of the King). Almost every middle film in a trilogy struggles with the structural problem of being narratively dependent on the films around it.
What Jackson and his co-writers (Fran Walsh, Philippa Boyens, Stephen Sinclair) did was reorganise Tolkien's source material to give the film its own three-act shape. The novel's structure — Frodo and Sam's storyline runs in parallel with the rest of the story across many chapters — was reshaped so that the film could deliver Helm's Deep as a clear third-act climax. The Frodo-Sam thread is structured around the gradual seduction of Sméagol back toward Gollum. The Aragorn thread is structured around Théoden's restoration. The Merry-Pippin thread is structured around the Ents' march to Isengard.
The result is the rare middle film that works as a standalone narrative experience. It is also, by general critical consensus, the strongest of the three. The trilogy's middle film is the most-formally-disciplined of the three.
The Battle of Helm's Deep occupies roughly the final 45 minutes of the film. The sequence required two months of shooting in New Zealand, with rain machines running through cold nights and roughly 200 stunt performers cycling through. The Uruk-hai forces visible on screen are produced through the Massive software platform that Weta Digital developed specifically for the trilogy — an AI-driven crowd-simulation system that gave each individual digital soldier their own behavioural rules.
The sequence is, on close inspection, structurally clever. The night-time setting is partly a creative choice (the dread of the unfamiliar dark) and partly a practical one (the digital army is more believable in lower-contrast lighting). The rain serves the same function. The arrival of Gandalf and the Riders of Rohan at dawn is timed for maximum visual contrast — the white-lit sunrise on the green field cutting through the night-and-rain palette of the preceding hour.
Gollum's animation in The Two Towers is, by general critical consensus, the foundational text of motion-capture performance in mainstream cinema. Andy Serkis performed every Gollum scene on set, in a motion-capture suit, with the other actors responding to him directly. The Weta animators then translated his performance into the digital character, frame by frame, with attention to micro-facial expressions and physical comportment that the digital pipeline had not previously preserved at this level.
The Academy's continued refusal to nominate Serkis for an acting Oscar — for Gollum, then for Kong, then for Caesar in the Planet of the Apes prequels — is widely considered one of the more-debated standing positions in Academy history. The structural argument is that motion-capture performance is acting in a way that fully-CGI character voice work is not. Serkis is doing the performance; the animators are translating it. The argument for nominating him has, across two decades, not yet persuaded the Academy.