We present an analysis of the merging cluster MACS J1149.5+2223 using archival imaging from Subaru/Suprime-Cam and multi-object spectroscopy from Keck/DEIMOS and Gemini/GMOS .
We employ two and three dimensional substructure tests and determine that MACS J1149.5+2223 is composed of two separate mergers between three subclusters occurring \sim 1 Gyr apart .
The primary merger gives rise to elongated X-ray morphology and a radio relic in the southeast .
The brightest cluster galaxy is a member of the northern subcluster of the primary merger .
This subcluster is very massive ( 16.7 ^ { + \text { 1.25 } } _ { - \text { 1.60 } } \times \text { 10 } ^ { \text { 14 } } M _ { \odot } ) .
The southern subcluster is also very massive ( 10.8 ^ { + \text { 3.37 } } _ { - \text { 3.54 } } \times \text { 10 } ^ { \text { 14 } } M _ { \odot } ) , yet it lacks an associated X-ray surface brightness peak , and it has been unidentified previously despite the detailed study of this Frontier Field cluster .
A secondary merger is occurring in the north along the line of sight with a third , less massive , subcluster ( 1.20 ^ { + \text { 0.19 } } _ { - \text { 0.34 } } \times \text { 10 } ^ { \text { 14 } } M _ { \odot } ) .
We perform a Monte Carlo dynamical analysis on the main merger and estimate a collision speed at pericenter of 2770 ^ { + \text { 610 } } _ { - \text { 310 } } km s ^ { - \text { 1 } } .
We show the merger to be returning from apocenter with core passage occurring 1.16 ^ { + \text { 0.50 } } _ { - \text { 0.25 } } Gyr before the observed state .
We identify the line of sight merging subcluster in a strong lensing analysis in the literature and show that it is likely bound to MACS J1149 despite having reached an extreme collision velocity of \sim 4000 km s ^ { - \text { 1 } } .