Nostoc and Richelia belong to a group of heterocystous cyanobacteria and are unique within this group in forming intracellular symbioses with phototrophic hosts, the angiosperm Gunnera and the diatoms (algae) Rhizosolenia and Hemiaulus, respectively. The function of the cyanobiont is similar in the symbioses, namely providing fixed atmospheric nitrogen to their hosts; also the cyanobionts are contained in a host compartment, the symbiosome. The evolutionary timescale for the cyanobiont-endosymbiosis formation is in both instances about a parts per thousand 90 Ma. However, the potentials for further co-evolution of host and microsymbiont, are different. Nostoc is regarded as preyed upon by its host, while in the Richelia-Rhizosolenia symbiosis example the evolution towards a new type of permanent organelle is possible. It is proposed that symbiosis is ruled by divergent host strategies. In the case of Richelia-Rhizosolenia the evolution of a permanent symbiosis is linked to diatom hosts needing to carry the cyanobiont permanently, as it is not available free-living in the oceans. However, in the case of Nostoc/Gunnera, the host exploits an abundant cyanobacterial species. A model where the relative abundance of microsymbionts determines the nature of the symbiosis comes into view: If environmental ratios of host/microsymbiont are so that hosts are the dominating party, then the host has to carry the microsymbiont as luggage (vertical transmission). Likewise, if the ratio of microsymbiont is higher than host, than the host will prey on the microsymbiont (horizontal transmission). The article also discusses the retention of secondary plastids in dinoflagellates. We show that dinoflagellates are organisms that exemplify both types of strategies that is either preying or harbouring a permanent organelle. The difference from the cyanobacterial example is that only parts of the eukaryotic microsymbionts are kept, usually only the plastid. We emphasize that the dinoflagellates can obtain their plastids from various different organisms. The luggage theory offers an explanation to why some dinoflagellate species contain kleptoplastids, while others have permanent, secondary plastids and some have tertiary plastids.